| { | |
| "model_type": "complexity-deep", | |
| "architecture": "DeepForCausalLM", | |
| "architectures": [ | |
| "DeepForCausalLM" | |
| ], | |
| "version": "0.13.0", | |
| "vocab_size": 32000, | |
| "hidden_size": 2048, | |
| "intermediate_size": 5632, | |
| "num_hidden_layers": 24, | |
| "num_attention_heads": 16, | |
| "num_key_value_heads": 8, | |
| "max_position_embeddings": 2048, | |
| "rope_theta": 10000.0, | |
| "rms_norm_eps": 1e-06, | |
| "attention_dropout": 0.0, | |
| "hidden_act": "silu", | |
| "tie_word_embeddings": true, | |
| "initializer_range": 0.02, | |
| "pad_token_id": 1, | |
| "bos_token_id": 2, | |
| "eos_token_id": 0, | |
| "use_token_routed_mlp": true, | |
| "num_experts": 4, | |
| "use_qk_norm": true, | |
| "use_sdpa": true, | |
| "dynamics_dt": 0.1, | |
| "dynamics_controller_hidden": 64, | |
| "torch_dtype": "float16" | |
| } |