config.json

{
  "architectures": [
    "LongcatFlashNgramForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "meituan-longcat/LongCat-Flash-Lite--configuration_longcat_ngram.LongcatFlashNgramConfig",
    "AutoModel": "meituan-longcat/LongCat-Flash-Lite--modeling_longcat_ngram.LongcatFlashNgramModel",
    "AutoModelForCausalLM": "meituan-longcat/LongCat-Flash-Lite--modeling_longcat_ngram.LongcatFlashNgramForCausalLM"
  },
  "bos_token_id": 1,
  "dtype": "bfloat16",
  "emb_neighbor_num": 2,
  "emb_split_num": 2,
  "eos_token_id": 2,
  "expert_ffn_hidden_size": 32,
  "ffn_hidden_size": 32,
  "head_dim": 192,
  "hidden_act": "silu",
  "hidden_size": 8,
  "initializer_range": 0.02,
  "kv_lora_rank": 384,
  "max_position_embeddings": 327680,
  "mla_scale_kv_lora": true,
  "mla_scale_q_lora": true,
  "model_type": "longcat_flash_ngram",
  "moe_topk": 12,
  "n_routed_experts": 32,
  "ngram_vocab_size_ratio": 4,
  "num_attention_heads": 4,
  "num_hidden_layers": 4,
  "num_key_value_heads": 4,
  "num_layers": 2,
  "q_lora_rank": 32,
  "qk_head_dim": 256,
  "qk_nope_head_dim": 64,
  "qk_rope_head_dim": 192,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "beta_fast": 32.0,
    "beta_slow": 1.0,
    "factor": 10.0,
    "mscale": 1,
    "mscale_all_dim": 1,
    "original_max_position_embeddings": 32768,
    "rope_type": "yarn"
  },
  "rope_theta": 5000000.0,
  "routed_scaling_factor": 6.0,
  "tie_word_embeddings": false,
  "transformers_version": "4.57.6",
  "use_cache": true,
  "v_head_dim": 64,
  "vocab_size": 131072,
  "zero_expert_num": 16,
  "zero_expert_type": "identity"
}