config.json

{
  "_name_or_path": "/code/syr/Llama-3.2-3B-Instruct",
  "architectures": [
    "LlavaLlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "audio_projector_type": "mlp2x_gelu",
  "bos_token_id": 128000,
  "eos_token_id": [
    128001,
    128008,
    128009
  ],
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 3072,
  "initializer_range": 0.02,
  "intermediate_size": 8192,
  "language": "English",
  "local_files_only": false,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "mm_audio_hidden_size": 1280,
  "mm_audio_select_feature": "patch",
  "mm_audio_select_layer": -1,
  "mm_audio_tower": "/code/syr/whisper-large-v3",
  "model_type": "llava_llama",
  "num_attention_heads": 24,
  "num_hidden_layers": 28,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "factor": 32.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "task": "transcribe",
  "tie_word_embeddings": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.43.4",
  "use_cache": false,
  "use_mm_proj": true,
  "vocab_size": 128261
}