AX-EmbeddingGemma-2-MLX-AXQ-MXFP4 / axquant_quantizer_execution.json
AutomatosX's picture
Publish AX-EmbeddingGemma-2-MLX-AXQ-MXFP4
22de0a9 verified
Raw History Blame Contribute Delete
55.9 kB
{
"created_at": "2026-10-11T15:22:59.320980Z",
"plan_sha256": "2c17cdda260f1d3706871f0f9abce2e403fe7e1982774ab1661f1fd6b216eb6a",
"records": [
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.embed_tokens",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.0.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.0.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.0.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.0.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.0.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.0.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.0.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.0.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.0.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.1.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.1.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.1.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.1.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.1.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.1.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.1.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.1.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.1.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.10.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.10.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.10.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.10.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.10.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.10.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.10.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.10.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.10.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.11.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.11.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.11.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.11.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.11.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.11.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.11.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.11.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.11.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.12.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.12.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.12.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.12.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.12.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.12.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.12.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.12.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.12.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.13.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.13.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.13.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.13.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.13.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.13.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.13.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.13.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.13.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.14.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.14.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.14.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.14.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.14.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.14.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.14.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.14.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.14.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.15.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.15.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.15.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.15.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.15.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.15.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.15.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.15.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.15.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.16.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.16.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.16.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.16.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.16.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.16.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.16.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.16.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.16.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.17.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.17.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.17.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.17.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.17.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.17.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.17.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.17.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.17.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.18.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.18.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.18.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.18.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.18.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.18.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.18.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.18.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.18.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.19.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.19.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.19.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.19.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.19.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.19.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.19.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.19.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.19.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.2.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.2.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.2.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.2.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.2.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.2.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.2.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.2.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.2.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.20.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.20.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.20.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.20.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.20.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.20.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.20.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.20.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.20.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.21.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.21.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.21.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.21.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.21.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.21.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.21.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.21.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.21.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.22.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.22.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.22.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.22.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.22.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.22.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.22.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.22.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.22.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.23.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.23.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.23.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.23.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.23.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.23.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.23.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.23.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.23.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.3.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.3.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.3.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.3.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.3.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.3.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.3.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.3.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.3.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.4.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.4.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.4.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.4.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.4.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.4.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.4.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.4.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.4.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.5.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.5.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.5.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.5.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.5.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.5.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.5.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.5.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.5.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.6.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.6.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.6.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.6.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.6.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.6.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.6.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.6.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.6.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.7.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.7.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.7.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.7.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.7.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.7.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.7.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.7.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.7.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.8.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.8.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.8.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.8.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.8.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.8.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.8.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.8.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.8.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.9.mlp.down_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.9.mlp.gate_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.9.mlp.up_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.9.ple_block.per_layer_input_gate",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.layers.9.ple_block.per_layer_projection",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.9.self_attn.k_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.9.self_attn.o_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.9.self_attn.q_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 4,
"fallback": false,
"group_size": 32,
"metadata": {},
"method": "mxfp4",
"module_path": "language_model.layers.9.self_attn.v_proj",
"note": "mlx-vlm affine packing",
"success": true
},
{
"bits": 8,
"fallback": false,
"group_size": 64,
"metadata": {},
"method": "affine",
"module_path": "language_model.ple.per_layer_model_projection",
"note": "mlx-vlm affine packing",
"success": true
}
],
"schema_version": "axquant.quantizer-execution.v2"
}