| { | |
| "architectures": [ | |
| "LongcatFlashForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "dtype": "float32", | |
| "eos_token_id": 2, | |
| "expert_ffn_hidden_size": 48, | |
| "ffn_hidden_size": 288, | |
| "head_dim": 4, | |
| "hidden_act": "silu", | |
| "hidden_size": 144, | |
| "initializer_range": 0.02, | |
| "is_decoder": true, | |
| "kv_lora_rank": 16, | |
| "max_position_embeddings": 200, | |
| "model_type": "longcat_flash", | |
| "moe_topk": 2, | |
| "n_routed_experts": 4, | |
| "num_attention_heads": 8, | |
| "num_hidden_layers": 2, | |
| "num_key_value_heads": 8, | |
| "num_layers": 1, | |
| "pad_token_id": 3, | |
| "q_lora_rank": 48, | |
| "qk_head_dim": 12, | |
| "qk_nope_head_dim": 8, | |
| "qk_rope_head_dim": 4, | |
| "rms_norm_eps": 1e-06, | |
| "rope_parameters": { | |
| "rope_theta": 10000000.0, | |
| "rope_type": "default" | |
| }, | |
| "routed_scaling_factor": 1.0, | |
| "tie_word_embeddings": false, | |
| "transformers_version": "5.16.0.dev0", | |
| "use_cache": true, | |
| "v_head_dim": 8, | |
| "vocab_size": 131072, | |
| "zero_expert_num": 2 | |
| } | |