{
  "_name_or_path": "/mnt/disk2/home/wujianfeng/LLaMA-Factory_audio/qwen2_mm_14B_pt_1/checkpoint-12000/",
  "architectures": [
    "Qwen2MMForConditionalGeneration"
  ],
  "attention_dropout": 0.0,
  "audio_config": {
    "_attn_implementation_autoset": true,
    "activation_dropout": 0.0,
    "activation_function": "relu",
    "architectures": [
      "SeamlessM4Tv2Model"
    ],
    "attention_dropout": 0.1,
    "dropout": 0.1,
    "encoder_attention_heads": 16,
    "encoder_ffn_dim": 8192,
    "encoder_layerdrop": 0.05,
    "encoder_layers": 24,
    "hidden_size": 1024,
    "initializer_range": 0.02,
    "is_encoder_decoder": true,
    "layer_norm_eps": 1e-05,
    "leaky_relu_slope": 0.1,
    "max_new_tokens": 256,
    "max_position_embeddings": 4096,
    "model_type": "qwen2_seamless_encoder",
    "num_attention_heads": 16,
    "num_hidden_layers": 24,
    "pad_token_id": 0,
    "resblock_dilation_sizes": [
      [
        1,
        3,
        5
      ],
      [
        1,
        3,
        5
      ],
      [
        1,
        3,
        5
      ]
    ],
    "resblock_kernel_sizes": [
      3,
      7,
      11
    ],
    "sampling_rate": 16000,
    "scale_embedding": true,
    "spkr_embed_dim": 256,
    "use_cache": true
  },
  "audio_end_token_id": 151667,
  "audio_start_token_id": 151666,
  "audio_token_id": 151665,
  "auto_map": {
    "AutoConfig": "configuration_qwen2_mm.Qwen2MMConfig",
    "AutoModel": "modeling_qwen2_mm.Qwen2MMForConditionalGeneration",
    "AutoModelForCausalLM": "modeling_qwen2_mm.Qwen2MMForConditionalGeneration"
  },
  "bos_token_id": 151643,
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 5120,
  "image_token_id": 151655,
  "initializer_range": 0.02,
  "intermediate_size": 13824,
  "llm_path": "/mnt/diskhd/Backup/DownloadModel/Qwen2.5-14B-Instruct/",
  "max_position_embeddings": 32768,
  "max_window_layers": 70,
  "model_type": "qwen2_mm",
  "num_attention_heads": 40,
  "num_hidden_layers": 48,
  "num_key_value_heads": 8,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000000.0,
  "sliding_window": 131072,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.45.0",
  "use_cache": false,
  "use_sliding_window": false,
  "video_token_id": 151656,
  "vision_config": {
    "model_type": "qwen2_vl"
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 152064
}