tiny-random/longcat-flash

1{2  "architectures": [3    "LongcatFlashForCausalLM"4  ],5  "attention_bias": false,6  "attention_dropout": 0.0,7  "attention_method": "MLA",8  "auto_map": {9    "AutoConfig": "configuration_longcat_flash.LongcatFlashConfig",10    "AutoModel": "modeling_longcat_flash.LongcatFlashModel",11    "AutoModelForCausalLM": "modeling_longcat_flash.LongcatFlashForCausalLM"12  },13  "bos_token_id": 1,14  "eos_token_id": 2,15  "expert_ffn_hidden_size": 64,16  "ffn_hidden_size": 64,17  "head_dim": 192,18  "hidden_act": "silu",19  "hidden_size": 8,20  "initializer_range": 0.006,21  "kv_lora_rank": 384,22  "max_position_embeddings": 131072,23  "mla_scale_kv_lora": true,24  "mla_scale_q_lora": true,25  "model_type": "longcat_flash",26  "moe_topk": 12,27  "n_routed_experts": 32,28  "norm_topk_prob": false,29  "num_attention_heads": 4,30  "num_key_value_heads": 4,31  "num_layers": 2,32  "q_lora_rank": 32,33  "qk_head_dim": 256,34  "qk_nope_head_dim": 64,35  "qk_rope_head_dim": 192,36  "rms_norm_eps": 1e-05,37  "rope_theta": 10000000.0,38  "routed_scaling_factor": 6.0,39  "router_bias": false,40  "tie_word_embeddings": false,41  "torch_dtype": "bfloat16",42  "transformers_version": "4.56.0.dev0",43  "use_cache": true,44  "v_head_dim": 64,45  "vocab_size": 131072,46  "zero_expert_num": 16,47  "zero_expert_type": "identity"48}