{ "architectures": [ "DeepseekV3ForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 1, "dtype": "bfloat16", "eos_token_id": 2, "first_k_dense_replace": 0, "head_dim": 4, "hidden_act": "silu", "hidden_size": 8, "initializer_range": 0.02, "intermediate_size": 16, "kv_lora_rank": 4, "max_position_embeddings": 1024, "model_type": "deepseek_v3", "moe_intermediate_size": 2048, "n_group": 1, "n_routed_experts": 2, "n_shared_experts": 1, "norm_topk_prob": true, "num_attention_heads": 2, "num_experts_per_tok": 1, "num_hidden_layers": 2, "num_key_value_heads": 2, "pad_token_id": 0, "pretraining_tp": 1, "q_lora_rank": 8, "qk_head_dim": 8, "qk_nope_head_dim": 4, "qk_rope_head_dim": 4, "rms_norm_eps": 1e-06, "rope_interleave": true, "rope_scaling": null, "rope_theta": 10000.0, "routed_scaling_factor": 1.0, "tie_word_embeddings": false, "topk_group": 1, "transformers_version": "4.56.0", "use_cache": true, "v_head_dim": 4, "vocab_size": 129280 }