| { | |
| "architectures": [ | |
| "DeepseekV4ForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 0, | |
| "eos_token_id": 1, | |
| "expert_dtype": "fp4", | |
| "head_dim": 16, | |
| "hidden_act": "silu", | |
| "hidden_size": 64, | |
| "initializer_range": 0.02, | |
| "max_position_embeddings": 256, | |
| "model_type": "deepseek_v4", | |
| "moe_intermediate_size": 32, | |
| "n_routed_experts": 8, | |
| "n_shared_experts": 1, | |
| "norm_topk_prob": true, | |
| "num_attention_heads": 4, | |
| "num_experts_per_tok": 2, | |
| "num_hash_layers": 0, | |
| "num_hidden_layers": 4, | |
| "num_key_value_heads": 1, | |
| "num_nextn_predict_layers": 0, | |
| "o_groups": 4, | |
| "o_lora_rank": 32, | |
| "q_lora_rank": 32, | |
| "qk_rope_head_dim": 8, | |
| "rms_norm_eps": 1e-20, | |
| "rope_scaling": { | |
| "beta_fast": 32, | |
| "beta_slow": 1, | |
| "factor": 16, | |
| "original_max_position_embeddings": 65536, | |
| "type": "yarn" | |
| }, | |
| "rope_theta": 10000, | |
| "routed_scaling_factor": 1.5, | |
| "scoring_func": "sqrtsoftplus", | |
| "sliding_window": 128, | |
| "swiglu_limit": 10.0, | |
| "tie_word_embeddings": false, | |
| "topk_method": "noaux_tc", | |
| "torch_dtype": "float32", | |
| "use_cache": true, | |
| "vocab_size": 256 | |
| } | |