config.json

{
  "_name_or_path": "liuhaotian/llava-v1.5-7b",
  "anchor_positions": null,
  "architectures": [
    "LlavaLlamaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "attn_fuse_hidden_act": "silu",
  "attn_fuse_num_heads": 4,
  "attn_fuse_size": 256,
  "attn_fuse_type": "AttnFuserV1",
  "bos_token_id": 1,
  "deep_supervision": false,
  "eos_token_id": 2,
  "freeze_mm_mlp_adapter": false,
  "freeze_mm_vision_resampler": false,
  "hidden_act": "silu",
  "hidden_size": 4096,
  "image_aspect_ratio": "pad",
  "initializer_range": 0.02,
  "intermediate_size": 11008,
  "le_dropout_prob": 0.1,
  "le_layers": [
    0,
    1,
    2,
    3,
    4,
    5,
    6,
    7,
    8,
    9,
    10,
    11,
    12,
    13,
    14,
    15,
    16,
    17,
    18,
    19,
    20,
    21,
    22,
    23,
    24,
    25,
    26,
    27,
    28,
    29,
    30,
    31
  ],
  "le_length": 1,
  "max_length": 4096,
  "max_position_embeddings": 4096,
  "max_remain_ratio": null,
  "min_remain_num": null,
  "mm_hidden_size": 1024,
  "mm_projector_type": "mlp2x_gelu",
  "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-patch14-336",
  "model_type": "llava_llama_gp",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "num_key_value_heads": 32,
  "ori_attn_supervision": false,
  "pad_token_id": 0,
  "pretraining_tp": 1,
  "reduce_layer": 21,
  "reduce_threshold": 0.5,
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_theta": 10000.0,
  "selected_layers": [
    21
  ],
  "selected_visual_layers": [
    23,
    17,
    11,
    5
  ],
  "tie_word_embeddings": false,
  "tokenizer_padding_side": "left",
  "torch_dtype": "float16",
  "transformers_version": "4.37.2",
  "tune_mm_mlp_adapter": false,
  "tune_mm_vision_resampler": false,
  "unfreeze_mm_vision_tower": false,
  "use_attention_logits": false,
  "use_cache": true,
  "use_mm_proj": true,
  "use_ref_masks": false,
  "use_zero_masks": false,
  "visual_cond_size": 512,
  "vocab_size": 32000
}