config.json

{
  "architectures": [
    "VipLlavaForConditionalGeneration"
  ],
  "ignore_index": -100,
  "image_token_index": 32000,
  "model_type": "vipllava",
  "pad_token_id": 32001,
  "projector_hidden_act": "gelu",
  "projector_layernorm_eps": 1e-05,
  "text_config": {
    "_name_or_path": "meta-llama/Llama-2-13b-hf",
    "architectures": [
      "LlamaForCausalLM"
    ],
    "hidden_size": 5120,
    "intermediate_size": 13824,
    "max_position_embeddings": 4096,
    "model_type": "llama",
    "num_attention_heads": 40,
    "num_hidden_layers": 40,
    "num_key_value_heads": 40,
    "rms_norm_eps": 1e-05,
    "torch_dtype": "float16",
    "vocab_size": 32064
  },
  "torch_dtype": "float16",
  "transformers_version": "4.37.0.dev0",
  "vision_config": {
    "hidden_size": 1024,
    "image_size": 336,
    "intermediate_size": 4096,
    "model_type": "clip_vision_model",
    "num_attention_heads": 16,
    "num_hidden_layers": 24,
    "patch_size": 14,
    "projection_dim": 768,
    "vocab_size": 32000
  },
  "vision_feature_layers": [
    -2,
    -5,
    -8,
    -11,
    6
  ],
  "vocab_size": 32064
}