config.json

{
  "architectures": [
    "MGMOmniForCausalLM"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "check_data_modality": false,
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 5120,
  "image_aspect_ratio": "qwen2vl",
  "image_token_id": 151655,
  "initializer_range": 0.02,
  "intermediate_size": 27648,
  "max_position_embeddings": 128000,
  "max_window_layers": 64,
  "mm_hidden_size": 1280,
  "mm_projector_type": "identity",
  "mm_speech_hidden_size": 1280,
  "mm_speech_projector_type": "dual_mlp_attn",
  "mm_speech_tower": "wcy1122/Dual-Qwen2A-7B-Encoder",
  "mm_use_im_patch_token": false,
  "mm_use_im_start_end": false,
  "mm_use_mrope": false,
  "mm_vision_select_feature": "patch",
  "mm_vision_select_layer": -1,
  "mm_vision_tower": "model_zoo/mgm_omni/Qwen2.5-VL-32B-ViT",
  "model_type": "MGMOmni",
  "num_attention_heads": 40,
  "num_hidden_layers": 64,
  "num_key_value_heads": 8,
  "overlap_ratio": 0.0,
  "pad_token_id": 151643,
  "rms_norm_eps": 1e-06,
  "rope_scaling": {
    "mrope_section": [
      16,
      24,
      24
    ],
    "rope_type": "default",
    "type": "default"
  },
  "rope_theta": 1000000.0,
  "sliding_window": 32768,
  "speech_encoder_ds_rate": 2,
  "speech_encoder_hidden_size": 2560,
  "speech_learn": false,
  "speech_norm": false,
  "speech_overlap": false,
  "speech_overlap_time": 30,
  "speech_tower_aux": "BELLE-2/Belle-whisper-large-v3-zh",
  "tie_word_embeddings": false,
  "tokenizer_model_max_length": 65536,
  "tokenizer_padding_side": "right",
  "torch_dtype": "float16",
  "train_modality": "text_image_speech",
  "transformers_version": "4.52.3",
  "use_cache": true,
  "use_mm_proj": true,
  "use_sliding_window": false,
  "video_token_id": 151656,
  "vision_config": {
    "depth": 32,
    "fullatt_block_indexes": [
      7,
      15,
      23,
      31
    ],
    "hidden_act": "silu",
    "hidden_size": 1280,
    "in_channels": 3,
    "in_chans": 3,
    "initializer_range": 0.02,
    "intermediate_size": 3456,
    "model_type": "qwen2_5_vl",
    "num_heads": 16,
    "out_hidden_size": 5120,
    "patch_size": 14,
    "spatial_merge_size": 2,
    "spatial_patch_size": 14,
    "temporal_patch_size": 2,
    "tokens_per_second": 2,
    "torch_dtype": "float16",
    "window_size": 112
  },
  "vision_end_token_id": 151653,
  "vision_start_token_id": 151652,
  "vision_token_id": 151654,
  "vocab_size": 152064,
  "weight_lambda": 0.0
}