{
  "anyres": false,
  "architectures": [
    "HyperCLOVAXOmniForCausalLM"
  ],
  "audio_config": {
    "activation_dropout": 0.0,
    "activation_function": "gelu",
    "architectures": [
      "Qwen2AudioEncoder"
    ],
    "attention_dropout": 0.0,
    "d_model": 1280,
    "dropout": 0.0,
    "encoder_attention_heads": 20,
    "encoder_ffn_dim": 5120,
    "encoder_layerdrop": 0.0,
    "encoder_layers": 32,
    "init_std": 0.02,
    "initializer_range": 0.02,
    "max_source_positions": 1500,
    "model_type": "qwen2_audio_encoder",
    "num_hidden_layers": 32,
    "num_mel_bins": 128,
    "scale_embedding": false
  },
  "audio_end_token_id": 128257,
  "audio_projector_type": "mlp",  
  "audio_start_token_id": 128256,
  "audio_token_id": 128071,
  "auto_map": {
    "AutoConfig": "configuration_hyperclovax_omni.HyperCLOVAXOmniConfig",
    "AutoModelForCausalLM": "modeling_hyperclovax_omni.HyperCLOVAXOmniForCausalLM",
    "AutoModelForSequenceClassification": "modeling_hyperclovax_omni.HyperCLOVAXOmniForSequenceClassification"
  },
  "discrete_audio_config": {
    "architectures": [
      "CosyvoiceEncoder"
    ],
    "auto_map": {
      "AutoConfig": "configuration_cosyvoice2.CosyVoice2Config",
      "AutoModel": "modeling_cosyvoice2.CosyVoice2Model"
    },
    "model_type": "cosyvoice2",
    "n_audio_ctx": 1500,
    "n_audio_head": 20,
    "n_audio_layer": 6,
    "n_audio_state": 1280,
    "n_codebook_size": 6561,
    "n_mels": 128,
    "use_sdpa": true
  },
  "discrete_audio_end_token_id": 128073,
  "discrete_audio_start_token_id": 128072,
  "discrete_audio_token_id": 128074,
  "discrete_audio_unit_0_id": 128606,
  "discrete_image_end_token_id": 128261,
  "discrete_image_start_token_id": 128260,
  "discrete_image_token_id": 128069,
  "discrete_image_unit_0_id": 135168,
  "discrete_vision_config": {
    "architectures": [
      "TextAlignedTokenizer"
    ],
    "auto_map": {
      "AutoConfig": "configuration_tatok.TATokConfig",
      "AutoModel": "modeling_tatok.TATokModel"
    },
    "bottleneck": {
      "args": {
        "bottleneck_dim": 1536,
        "norm": "none",
        "regularizer": {
          "args": {
            "codebook_loss_weight": 1.0,
            "codebook_size": 65536,
            "commitment_loss_weight": 0.25,
            "entropy_loss_temperature": 0.01,
            "entropy_loss_weight": 0.0,
            "l2_normalized": true,
            "residual_weight": 0.1,
            "stochastic": true,
            "stochastic_temperature": 0.03,
            "top_k": 4,
            "top_k_prob": 0.5
          },
          "name": "simvq"
        }
      },
      "name": "bottleneck"
    },
    "bottleneck_token_num": 729,
    "decoder_depth": 3,
    "input_size": 384,
    "input_type": "indices",
    "model_type": "tatok",
    "pool_scale": 1,
    "rand_scale": true,
    "select_layer_id": -2,
    "teacher": "google/siglip2-so400m-patch14-384"
  },
  "eos_token_id": 128001,
  "hidden_size": 4096,
  "ignore_index": -100,
  "image_end_token_id": 128058,
  "image_start_token_id": 128059,
  "image_token_id": 128062,
  "is_safetensor_save": true,
  "max_num_grids": -1,
  "mm_projector_type": "linear",
  "model_type": "hyperclovax_omni",
  "num_queries_vis_abstractor": -1,
  "num_queries_vis_abstractors": -1,
  "possible_resolutions": [],
  "proj_pos_emb": true,
  "proj_prenorm": false,
  "text_config": {
    "architectures": [
      "LlamaForCausalLM"
    ],
    "attention_bias": false,
    "attention_dropout": 0.0,
    "bos_token_id": 128000,
    "eos_token_id": 128001,
    "head_dim": 128,
    "hidden_act": "silu",
    "hidden_size": 4096,
    "initializer_range": 0.02,
    "intermediate_size": 12288,
    "logits_scaling": 1.0,
    "max_position_embeddings": 8192,
    "mlp_bias": false,
    "model_type": "llama",
    "num_attention_heads": 32,
    "num_hidden_layers": 36,
    "num_key_value_heads": 8,
    "pretraining_tp": 1,
    "rms_norm_eps": 1e-06,
    "rope_scaling": null,
    "rope_theta": 5000000,
    "tie_word_embeddings": false,
    "use_cache": true,
    "vocab_size": 200704
  },
  "torch_dtype": "float32",
  "unpad": false,
  "use_1x1_grid": false,
  "use_components": {
    "save_discrete_audio_model": true,
    "save_discrete_vision_model": true,
    "use_cont_audio": true,
    "use_discrete_audio": true,
    "use_discrete_vision": true
  },
  "vision_feature_layer": -2,
  "video_audio_compressor_config": {
    "architectures": null,
    "auto_map": {
      "AutoConfig": "configuration_mambamia.MambaMiaVideoAudioCompressorConfig",
      "AutoModel": "modeling_mambamia.MambaMiaVideoAudioCompressor"
    },
    "chunk_size": 25,
    "hidden_size": 3072,
    "input_size": 4096,
    "model_type": "mambamia_videoaudio_compressor",
    "num_hidden_layers": 1,
    "output_size": 4096
  },
  "video_audio_compressor_type": "mambamia",
  "video_audio_token_id": 128070,
  "video_end_token_id": 128061,
  "video_first_last_frames_slows": null,
  "video_num_queries_fast": null,
  "video_num_queries_slow": null,
  "video_start_token_id": 128060,
  "video_token_id": 128063,
  "vision_config": {
    "anyres": false,
    "architectures": [
      "Qwen2_5_VisionTransformerPretrainedModel"
    ],
    "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,
    "max_num_grids": -1,
    "model_type": "qwen2_5_vl_visual",
    "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,
    "window_size": 112
  },
  "vision_input_chunk_size": null
}