{
  "architectures": [
    "NemotronOmniForConditionalGeneration"
  ],
  "initializer_range": 0.02,
  
  "text_config": {
    "attention_bias": false,
    "model_type": "llama_text",
    "attention_dropout": 0.0,
    "bos_token_id": 128000,
    "eos_token_id": 128001,
    "head_dim": 128,
    "hidden_act": "silu",
    "hidden_size": 3072,
    "intermediate_size": 9216,
    "max_position_embeddings": 131072,
    "mlp_bias": false,
    "num_attention_heads": 32,
    "num_hidden_layers": 32,
    "num_key_value_heads": 8,
    "pretraining_tp": 1,
    "rms_norm_eps": 1e-05,
    "rope_scaling": {
      "factor": 4.0,
      "low_freq_factor": 1.0,
      "high_freq_factor": 4.0,
      "original_max_position_embeddings": 8192,
      "rope_type": "llama3"
    },
    "rope_theta": 3565775107.2609234,
    "tie_word_embeddings": false,
    "torch_dtype": "bfloat16",
    "transformers_version": "4.47.1",
    "use_cache": true,
    "vocab_size": 128256
  },
  
  "audio_config": {
    "activation": "swish",
    "activation_checkpointing": {
      "interval": 1,
      "module": "transformer",
      "offload": false
    },
    "attention_dim": 1024,
    "attention_heads": 16,
    "batch_norm": false,
    "bias_in_glu": true,
    "causal": true,
    "chunk_size": -1,
    "cnn_layer_norm": true,
    "conv_activation": "swish",
    "conv_glu_type": "swish",
    "depthwise_multiplier": 1,
    "depthwise_seperable_out_channel": 1024,
    "dropout_rate": 0.0,
    "encoder_embedding_config": {
      "input_size": 80
    },
    "ext_pw_kernel_size": 1,
    "ext_pw_out_channel": 1024,
    "input_layer": "nemo_conv",
    "input_size": 80,
    "kernel_size": 3,
    "left_chunk": 18,
    "linear_units": 1536,
    "nemo_conv_settings": {
      "conv_channels": 1024
    },
    "num_blocks": 24,
    "relative_attention_bias_args": {
      "t5_bias_max_distance": 500,
      "type": "t5"
    },
    "time_reduction": 8
  },
  "speech_lora": {
    "dp": 0.01,
    "layer": "((layers.*self_attn\\.(q|k|v|o)_proj)|(layers.*mlp\\.(gate|up|down)_proj))",
    "lora_alpha": 320,
    "r": 320,
    "use_rslora": true
  },
  "auto_map": {
    "AutoConfig": "configuration_nemotronOmni.NemotronOmniConfig",
    "AutoModel": "modeling_NemotronOmni.NemotronOmniForConditionalGeneration"
  },
  "audio_token_index": 128255,
  "boa_token_index": 128253,
  "eoa_token_index": 128254,
  "eos_token_id": [
    128001,
    128008,
    128009
  ],
  "model_type": "NemotronOmni"
}