| { | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "architectures": [ | |
| "HunYuanForCausalLM" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_hunyuan.HunYuanConfig", | |
| "AutoModel": "modeling_hunyuan.HunyuanModel", | |
| "AutoModelForCausalLM": "modeling_hunyuan.HunYuanForCausalLM" | |
| }, | |
| "bos_token_id": 1, | |
| "capacity_factor": 1.0, | |
| "cla_share_factor": 2, | |
| "eos_token_id": 2, | |
| "hidden_act": "silu", | |
| "hidden_size": 6400, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 18304, | |
| "max_position_embeddings": 262144, | |
| "model_type": "hunyuan", | |
| "moe_drop_tokens": false, | |
| "moe_random_routing_dropped_token": false, | |
| "moe_topk": 1, | |
| "num_attention_heads": 80, | |
| "num_experts": 16, | |
| "num_hidden_layers": 64, | |
| "num_key_value_heads": 8, | |
| "num_shared_expert": 1, | |
| "pad_token_id": 0, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": { | |
| "alpha": 100000.0, | |
| "factor": 1.0, | |
| "type": "dynamic" | |
| }, | |
| "rope_theta": 10000.0, | |
| "tie_word_embeddings": true, | |
| "transformers_version": "4.41.2", | |
| "use_cache": true, | |
| "use_cla": true, | |
| "use_mixed_mlp_moe": true, | |
| "use_qk_norm": true, | |
| "vocab_size": 128512, | |
| "torch_dtype": "bfloat16" | |
| } | |