| { | |
| "_flash_attn_2_enabled": true, | |
| "architectures": [ | |
| "MixtralForCausalLM" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_mixtral_clex.CLEXMixtralConfig", | |
| "AutoModelForCausalLM": "modeling_mixtral_clex.MixtralForCausalLM" | |
| }, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "hidden_act": "silu", | |
| "hidden_size": 4096, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 14336, | |
| "log_scale": false, | |
| "max_position_embeddings": 32768, | |
| "model_type": "mixtral", | |
| "num_attention_heads": 32, | |
| "num_experts_per_tok": 2, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 8, | |
| "num_local_experts": 8, | |
| "output_router_logits": false, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": { | |
| "factor": 1, | |
| "max_factor": 10, | |
| "param_factor": 2, | |
| "time_dt": 0.01, | |
| "type": "clex", | |
| "act": "tanh" | |
| }, | |
| "rope_theta": 1000000.0, | |
| "router_aux_loss_coef": 0.02, | |
| "sliding_window": null, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.36.2", | |
| "use_cache": true, | |
| "use_flashattn": true, | |
| "vocab_size": 32000 | |
| } | |