PrimeIntellect/Qwen3-30B-A3B-Base-Fast

1{2  "architectures": [3    "Qwen3MoeForCausalLM"4  ],5  "auto_map": {6    "AutoConfig": "configuration_qwen3_moe.Qwen3MoeConfig",7    "AutoModelForCausalLM": "modeling_qwen3_moe.Qwen3MoeForCausalLM",8    "AutoModel": "modeling_qwen3_moe.Qwen3MoeModel"9  },10  "attention_bias": false,11  "attention_dropout": 0.0,12  "bos_token_id": 151643,13  "decoder_sparse_step": 1,14  "eos_token_id": 151645,15  "head_dim": 128,16  "hidden_act": "silu",17  "hidden_size": 2048,18  "initializer_range": 0.02,19  "intermediate_size": 6144,20  "max_position_embeddings": 40960,21  "max_window_layers": 48,22  "mlp_only_layers": [],23  "model_type": "qwen3_moe",24  "moe_intermediate_size": 768,25  "norm_topk_prob": true,26  "num_attention_heads": 32,27  "num_experts": 128,28  "num_experts_per_tok": 8,29  "num_hidden_layers": 48,30  "num_key_value_heads": 4,31  "output_router_logits": false,32  "rms_norm_eps": 1e-06,33  "rope_scaling": null,34  "rope_theta": 1000000.0,35  "router_aux_loss_coef": 0.001,36  "sliding_window": null,37  "tie_word_embeddings": false,38  "torch_dtype": "bfloat16",39  "transformers_version": "4.51.0",40  "use_cache": true,41  "use_sliding_window": false,42  "vocab_size": 15193643}