{
  "model_type": "bidirlm_omni",
  "transformers_version": "5.5.0",
  "auto_map": {
    "AutoConfig": "configuration_bidirlm_omni.BidirLMOmniConfig",
    "AutoModel": "modeling_bidirlm_omni.BidirLMOmniModel",
    "AutoModelForMaskedLM": "modeling_bidirlm_omni.BidirLMOmniForMaskedLM",
    "AutoModelForSequenceClassification": "modeling_bidirlm_omni.BidirLMOmniForSequenceClassification",
    "AutoModelForTokenClassification": "modeling_bidirlm_omni.BidirLMOmniForTokenClassification"
  },
  "trust_remote_code": true,
  "freeze_audio": true,
  "freeze_visual": true,
  "max_image_size": null,
  "max_sequence_length": 1024,
  "rope_scaling": {
    "rope_type": "default",
    "rope_theta": 5000000.0,
    "mrope_section": [
      24,
      20,
      20
    ]
  },
  "audio_config": {
    "model_type": "bidirlm_omni_audio",
    "num_mel_bins": 128,
    "encoder_layers": 24,
    "encoder_attention_heads": 16,
    "encoder_ffn_dim": 4096,
    "d_model": 1024,
    "downsample_hidden_size": 480,
    "output_dim": 2048,
    "max_source_positions": 1500,
    "n_window": 100,
    "n_window_infer": 400,
    "conv_chunksize": 500
  },
  "vision_config": {
    "model_type": "bidirlm_omni_vision",
    "depth": 24,
    "hidden_size": 1024,
    "intermediate_size": 4096,
    "num_heads": 16,
    "in_channels": 3,
    "patch_size": 16,
    "spatial_merge_size": 2,
    "temporal_patch_size": 2,
    "out_hidden_size": 2048,
    "num_position_embeddings": 2304,
    "deepstack_visual_indexes": [8, 16, 24]
  },
  "text_config": {
    "model_type": "bidirlm_omni_text",
    "vocab_size": 151936,
    "hidden_size": 2048,
    "intermediate_size": 6144,
    "num_hidden_layers": 28,
    "num_attention_heads": 16,
    "num_key_value_heads": 8,
    "head_dim": 128,
    "rms_norm_eps": 1e-6,
    "rope_theta": 5000000.0,
    "rope_scaling": {
      "rope_type": "default",
      "rope_theta": 5000000.0,
      "mrope_section": [24, 20, 20]
    },
    "tie_word_embeddings": false
  }
}