| { | |
| "architectures": [ | |
| "Qwen3MoeForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "auto_map": { | |
| "AutoConfig": "configuration_qwen3_moe.Qwen3MoeConfig", | |
| "AutoModel": "modeling_qwen3_moe.Qwen3MoeModel", | |
| "AutoModelForCausalLM": "modeling_qwen3_moe.Qwen3MoeForCausalLM" | |
| }, | |
| "bos_token_id": 151643, | |
| "decoder_sparse_step": 1, | |
| "dtype": "bfloat16", | |
| "eos_token_id": 151645, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_size": 1024, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 6144, | |
| "load_balance_coeff": 0.001, | |
| "max_position_embeddings": 4096, | |
| "max_window_layers": 48, | |
| "mlp_only_layers": [], | |
| "model_type": "qwen3_moe", | |
| "moe_intermediate_size": 256, | |
| "norm_topk_prob": true, | |
| "num_attention_heads": 16, | |
| "num_experts": 16, | |
| "num_experts_per_tok": 4, | |
| "num_hidden_layers": 12, | |
| "num_key_value_heads": 4, | |
| "output_router_logits": false, | |
| "rms_norm_eps": 1e-06, | |
| "rope_scaling": null, | |
| "rope_theta": 1000000.0, | |
| "router_aux_loss_coef": 0.001, | |
| "sliding_window": null, | |
| "tie_word_embeddings": false, | |
| "transformers_version": "4.56.1", | |
| "use_cache": false, | |
| "use_grouped_mm": true, | |
| "use_sliding_window": false, | |
| "vocab_size": 151936 | |
| } | |