| { |
| "architectures": [ |
| "LagunaForCausalLM" |
| ], |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "auto_map": { |
| "AutoConfig": "configuration_laguna.LagunaConfig", |
| "AutoModelForCausalLM": "modeling_laguna.LagunaForCausalLM" |
| }, |
| "bos_token_id": 2, |
| "eos_token_id": [ |
| 2, |
| 24 |
| ], |
| "gating": true, |
| "head_dim": 128, |
| "hidden_size": 2048, |
| "intermediate_size": 8192, |
| "layer_types": [ |
| "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" |
| ], |
| "max_position_embeddings": 131072, |
| "mlp_layer_types": [ |
| "dense", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse", |
| "sparse" |
| ], |
| "model_type": "laguna", |
| "moe_apply_router_weight_on_input": false, |
| "moe_intermediate_size": 512, |
| "moe_routed_scaling_factor": 2.5, |
| "num_attention_heads": 48, |
| "num_attention_heads_per_layer": [ |
| 48, |
| 64, |
| 64, |
| 64, |
| 48, |
| 64, |
| 64, |
| 64, |
| 48, |
| 64, |
| 64, |
| 64, |
| 48, |
| 64, |
| 64, |
| 64, |
| 48, |
| 64, |
| 64, |
| 64, |
| 48, |
| 64, |
| 64, |
| 64, |
| 48, |
| 64, |
| 64, |
| 64, |
| 48, |
| 64, |
| 64, |
| 64, |
| 48, |
| 64, |
| 64, |
| 64, |
| 48, |
| 64, |
| 64, |
| 64 |
| ], |
| "num_experts": 256, |
| "num_experts_per_tok": 8, |
| "num_hidden_layers": 40, |
| "num_key_value_heads": 8, |
| "pad_token_id": 9, |
| "partial_rotary_factor": 0.5, |
| "quantization": { |
| "group_size": 16, |
| "bits": 4, |
| "mode": "nvfp4", |
| "model.layers.1.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.2.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.3.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.4.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.5.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.6.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.7.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.8.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.9.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.10.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.11.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.12.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.13.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.14.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.15.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.16.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.17.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.18.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.19.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.20.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.21.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.22.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.23.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.24.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.25.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.26.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.27.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.28.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.29.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.30.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.31.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.32.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.33.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.34.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.35.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.36.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.37.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.38.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.39.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| } |
| }, |
| "quantization_config": { |
| "group_size": 16, |
| "bits": 4, |
| "mode": "nvfp4", |
| "model.layers.1.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.2.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.3.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.4.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.5.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.6.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.7.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.8.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.9.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.10.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.11.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.12.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.13.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.14.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.15.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.16.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.17.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.18.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.19.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.20.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.21.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.22.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.23.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.24.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.25.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.26.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.27.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.28.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.29.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.30.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.31.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.32.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.33.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.34.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.35.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.36.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.37.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.38.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| }, |
| "model.layers.39.mlp.gate.proj": { |
| "group_size": 64, |
| "bits": 8 |
| } |
| }, |
| "rms_norm_eps": 1e-06, |
| "rope_parameters": { |
| "full_attention": { |
| "rope_theta": 500000.0, |
| "rope_type": "yarn", |
| "factor": 32.0, |
| "original_max_position_embeddings": 4096, |
| "beta_slow": 1.0, |
| "beta_fast": 64.0, |
| "attention_factor": 1.0, |
| "partial_rotary_factor": 0.5 |
| }, |
| "sliding_attention": { |
| "rope_type": "default", |
| "rope_theta": 10000.0, |
| "partial_rotary_factor": 1.0 |
| }, |
| "original_max_position_embeddings": 4096 |
| }, |
| "router_aux_loss_coef": 0.0, |
| "shared_expert_intermediate_size": 512, |
| "sliding_window": 512, |
| "tie_word_embeddings": false, |
| "torch_dtype": "bfloat16", |
| "use_cache": true, |
| "vocab_size": 100352 |
| } |
