{
  "architectures": [
    "KimiK25ForConditionalGeneration"
  ],
  "model_type": "kimi_k25",
  "text_config": {
    "architectures": [
      "DeepseekV3ForCausalLM"
    ],
    "attention_bias": false,
    "attention_dropout": 0.0,
    "bos_token_id": 0,
    "dtype": "float32",
    "eos_token_id": 1,
    "first_k_dense_replace": 1,
    "head_dim": 64,
    "hidden_act": "silu",
    "hidden_size": 1024,
    "initializer_range": 0.02,
    "intermediate_size": 2048,
    "kv_lora_rank": 512,
    "max_position_embeddings": 4096,
    "model_type": "kimi_k2",
    "moe_intermediate_size": 256,
    "n_group": 1,
    "n_routed_experts": 8,
    "n_shared_experts": 1,
    "norm_topk_prob": true,
    "num_attention_heads": 16,
    "num_experts_per_tok": 4,
    "num_hidden_layers": 24,
    "num_key_value_heads": 16,
    "pretraining_tp": 1,
    "q_lora_rank": 512,
    "qk_head_dim": 192,
    "qk_nope_head_dim": 128,
    "qk_rope_head_dim": 64,
    "rms_norm_eps": 1e-06,
    "rope_interleave": true,
    "rope_scaling": null,
    "rope_theta": 50000,
    "routed_scaling_factor": 2.827,
    "scoring_func": "sigmoid",
    "tie_word_embeddings": false,
    "topk_group": 1,
    "topk_method": "noaux_tc",
    "transformers_version": "4.57.6",
    "use_cache": true,
    "use_grouped_mm": false,
    "v_head_dim": 128,
    "vocab_size": 129280
  },
  "vision_config": {
    "mm_hidden_size": 64,
    "mm_projector_type": "patchmerger",
    "patch_size": 14,
    "text_hidden_size": 1024,
    "vt_hidden_size": 64,
    "vt_intermediate_size": 128,
    "vt_num_attention_heads": 2,
    "vt_num_hidden_layers": 2
  },
  "tie_word_embeddings": false
}