| { |
| "architectures": [ |
| "FlintForCausalLM" |
| ], |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "bos_token_id": 150000, |
| "decoder_sparse_step": 1, |
| "eos_token_id": [ |
| 150001, |
| 150003, |
| 150005, |
| 150018 |
| ], |
| "head_dim": 128, |
| "hidden_act": "silu", |
| "hidden_size": 2048, |
| "initializer_range": 0.02, |
| "intermediate_size": 6144, |
| "layer_types": [ |
| "full_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention", |
| "sliding_attention", |
| "sliding_attention", |
| "full_attention" |
| ], |
| "max_position_embeddings": 262144, |
| "mlp_only_layers": [ |
| 0, |
| 1 |
| ], |
| "model_type": "flint", |
| "moe_intermediate_size": 512, |
| "no_rope_layer_indices": [ |
| 0, |
| 1, |
| 5, |
| 9, |
| 13, |
| 17, |
| 21, |
| 25, |
| 29, |
| 33, |
| 37, |
| 41, |
| 45, |
| 48 |
| ], |
| "norm_topk_prob": true, |
| "num_attention_heads": 32, |
| "num_experts": 256, |
| "num_experts_per_tok": 6, |
| "num_hidden_layers": 49, |
| "num_key_value_heads": 4, |
| "output_router_logits": false, |
| "pad_token_id": 150001, |
| "quantization_config": { |
| "config_groups": { |
| "group_0": { |
| "input_activations": { |
| "dynamic": false, |
| "group_size": 16, |
| "num_bits": 4, |
| "type": "float" |
| }, |
| "targets": [ |
| "Linear" |
| ], |
| "weights": { |
| "dynamic": false, |
| "group_size": 16, |
| "num_bits": 4, |
| "type": "float" |
| } |
| } |
| }, |
| "ignore": [ |
| "lm_head", |
| "model.embed_tokens", |
| "model.layers.0.*", |
| "model.layers.1.*", |
| "model.layers.2.self_attn*", |
| "model.layers.2.mlp.gate", |
| "model.layers.2.mlp.shared_expert*", |
| "model.layers.3.self_attn*", |
| "model.layers.3.mlp.gate", |
| "model.layers.3.mlp.shared_expert*", |
| "model.layers.4.self_attn*", |
| "model.layers.4.mlp.gate", |
| "model.layers.4.mlp.shared_expert*", |
| "model.layers.5.self_attn*", |
| "model.layers.5.mlp.gate", |
| "model.layers.5.mlp.shared_expert*", |
| "model.layers.6.self_attn*", |
| "model.layers.6.mlp.gate", |
| "model.layers.6.mlp.shared_expert*", |
| "model.layers.7.self_attn*", |
| "model.layers.7.mlp.gate", |
| "model.layers.7.mlp.shared_expert*", |
| "model.layers.8.self_attn*", |
| "model.layers.8.mlp.gate", |
| "model.layers.8.mlp.shared_expert*", |
| "model.layers.9.self_attn*", |
| "model.layers.9.mlp.gate", |
| "model.layers.9.mlp.shared_expert*", |
| "model.layers.10.self_attn*", |
| "model.layers.10.mlp.gate", |
| "model.layers.10.mlp.shared_expert*", |
| "model.layers.11.self_attn*", |
| "model.layers.11.mlp.gate", |
| "model.layers.11.mlp.shared_expert*", |
| "model.layers.12.self_attn*", |
| "model.layers.12.mlp.gate", |
| "model.layers.12.mlp.shared_expert*", |
| "model.layers.13.self_attn*", |
| "model.layers.13.mlp.gate", |
| "model.layers.13.mlp.shared_expert*", |
| "model.layers.14.self_attn*", |
| "model.layers.14.mlp.gate", |
| "model.layers.14.mlp.shared_expert*", |
| "model.layers.15.self_attn*", |
| "model.layers.15.mlp.gate", |
| "model.layers.15.mlp.shared_expert*", |
| "model.layers.16.self_attn*", |
| "model.layers.16.mlp.gate", |
| "model.layers.16.mlp.shared_expert*", |
| "model.layers.17.self_attn*", |
| "model.layers.17.mlp.gate", |
| "model.layers.17.mlp.shared_expert*", |
| "model.layers.18.self_attn*", |
| "model.layers.18.mlp.gate", |
| "model.layers.18.mlp.shared_expert*", |
| "model.layers.19.self_attn*", |
| "model.layers.19.mlp.gate", |
| "model.layers.19.mlp.shared_expert*", |
| "model.layers.20.self_attn*", |
| "model.layers.20.mlp.gate", |
| "model.layers.20.mlp.shared_expert*", |
| "model.layers.21.self_attn*", |
| "model.layers.21.mlp.gate", |
| "model.layers.21.mlp.shared_expert*", |
| "model.layers.22.self_attn*", |
| "model.layers.22.mlp.gate", |
| "model.layers.22.mlp.shared_expert*", |
| "model.layers.23.self_attn*", |
| "model.layers.23.mlp.gate", |
| "model.layers.23.mlp.shared_expert*", |
| "model.layers.24.self_attn*", |
| "model.layers.24.mlp.gate", |
| "model.layers.24.mlp.shared_expert*", |
| "model.layers.25.self_attn*", |
| "model.layers.25.mlp.gate", |
| "model.layers.25.mlp.shared_expert*", |
| "model.layers.26.self_attn*", |
| "model.layers.26.mlp.gate", |
| "model.layers.26.mlp.shared_expert*", |
| "model.layers.27.self_attn*", |
| "model.layers.27.mlp.gate", |
| "model.layers.27.mlp.shared_expert*", |
| "model.layers.28.self_attn*", |
| "model.layers.28.mlp.gate", |
| "model.layers.28.mlp.shared_expert*", |
| "model.layers.29.self_attn*", |
| "model.layers.29.mlp.gate", |
| "model.layers.29.mlp.shared_expert*", |
| "model.layers.30.self_attn*", |
| "model.layers.30.mlp.gate", |
| "model.layers.30.mlp.shared_expert*", |
| "model.layers.31.self_attn*", |
| "model.layers.31.mlp.gate", |
| "model.layers.31.mlp.shared_expert*", |
| "model.layers.32.self_attn*", |
| "model.layers.32.mlp.gate", |
| "model.layers.32.mlp.shared_expert*", |
| "model.layers.33.self_attn*", |
| "model.layers.33.mlp.gate", |
| "model.layers.33.mlp.shared_expert*", |
| "model.layers.34.self_attn*", |
| "model.layers.34.mlp.gate", |
| "model.layers.34.mlp.shared_expert*", |
| "model.layers.35.self_attn*", |
| "model.layers.35.mlp.gate", |
| "model.layers.35.mlp.shared_expert*", |
| "model.layers.36.self_attn*", |
| "model.layers.36.mlp.gate", |
| "model.layers.36.mlp.shared_expert*", |
| "model.layers.37.self_attn*", |
| "model.layers.37.mlp.gate", |
| "model.layers.37.mlp.shared_expert*", |
| "model.layers.38.self_attn*", |
| "model.layers.38.mlp.gate", |
| "model.layers.38.mlp.shared_expert*", |
| "model.layers.39.self_attn*", |
| "model.layers.39.mlp.gate", |
| "model.layers.39.mlp.shared_expert*", |
| "model.layers.40.self_attn*", |
| "model.layers.40.mlp.gate", |
| "model.layers.40.mlp.shared_expert*", |
| "model.layers.41.self_attn*", |
| "model.layers.41.mlp.gate", |
| "model.layers.41.mlp.shared_expert*", |
| "model.layers.42.self_attn*", |
| "model.layers.42.mlp.gate", |
| "model.layers.42.mlp.shared_expert*", |
| "model.layers.43.self_attn*", |
| "model.layers.43.mlp.gate", |
| "model.layers.43.mlp.shared_expert*", |
| "model.layers.44.self_attn*", |
| "model.layers.44.mlp.gate", |
| "model.layers.44.mlp.shared_expert*", |
| "model.layers.45.self_attn*", |
| "model.layers.45.mlp.gate", |
| "model.layers.45.mlp.shared_expert*", |
| "model.layers.46.self_attn*", |
| "model.layers.46.mlp.gate", |
| "model.layers.46.mlp.shared_expert*", |
| "model.layers.47.self_attn*", |
| "model.layers.47.mlp.gate", |
| "model.layers.47.mlp.shared_expert*", |
| "model.layers.48.self_attn*", |
| "model.layers.48.mlp.gate", |
| "model.layers.48.mlp.shared_expert*" |
| ], |
| "quant_algo": "NVFP4", |
| "quant_method": "modelopt" |
| }, |
| "rms_norm_eps": 1e-06, |
| "rope_theta": 10000.0, |
| "router_aux_loss_coef": 0.0005, |
| "router_score_function": "sigmoid", |
| "shared_expert_intermediate_size": 1024, |
| "sliding_window": 513, |
| "softmax_type": "learnable", |
| "tie_word_embeddings": false, |
| "torch_dtype": "bfloat16", |
| "transformers_version": "5.12.1", |
| "use_cache": true, |
| "use_sliding_window": true, |
| "vocab_size": 150144 |
| } |
|
|
