{
  "_name_or_path": "./temp",
  "architectures": [
    "LlavaForCausalLM"
  ],
  "initializer_factor": 1.0,
  "initializer_range": 0.02,
  "model_type": "llava",
  "text_config": {
    "max_position_embeddings": 4096,
    "model_type": "llama",
    "torch_dtype": "float16"
  },
  "torch_dtype": "bfloat16",
  "transformers_version": "4.36.0.dev0",
  "vision_config": {
    "bos_token_id": 1,
    "eos_token_id": 2,
    "freeze_mm_mlp_adapter": true,
    "hidden_act": "silu",
    "image_aspect_ratio": "square",
    "image_grid_pinpoints": null,
    "max_position_embeddings": 2048,
    "mm_resampler_type": null,
    "mm_use_im_patch_token": false,
    "mm_use_im_start_end": false,
    "mm_vision_select_feature": "patch",
    "mm_vision_select_layer": -2,
    "mm_vision_tower": "openai/clip-vit-large-patch14",
    "model_type": "llava_vision",
    "pad_token_id": 0,
    "pretraining_tp": 1,
    "projector": "None",
    "rms_norm_eps": 1e-06,
    "rope_scaling": null,
    "tie_word_embeddings": false,
    "torch_dtype": "float16",
    "tune_mm_mlp_adapter": false,
    "tune_mm_vision_resampler": false,
    "use_mm_proj": true
  }
}