{
  "_name_or_path": "google/vit-base-patch16-224-in21k",
  "architectures": [
    "ViTForImageClassification"
  ],
  "attention_probs_dropout_prob": 0.0,
  "encoder_stride": 16,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.0,
  "hidden_size": 768,
  "id2label": {
    "0": "banana custom",
    "1": "bicycle",
    "10": "smoke from fire",
    "11": "steam from man hole",
    "12": "street",
    "2": "big bird",
    "3": "car",
    "4": "carriage",
    "5": "flocks of insects",
    "6": "leaves",
    "7": "motorbyke",
    "8": "person",
    "9": "small bird"
  },
  "image_size": 224,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "banana custom": "0",
    "bicycle": "1",
    "big bird": "2",
    "car": "3",
    "carriage": "4",
    "flocks of insects": "5",
    "leaves": "6",
    "motorbyke": "7",
    "person": "8",
    "small bird": "9",
    "smoke from fire": "10",
    "steam from man hole": "11",
    "street": "12"
  },
  "layer_norm_eps": 1e-12,
  "model_type": "vit",
  "num_attention_heads": 12,
  "num_channels": 3,
  "num_hidden_layers": 12,
  "patch_size": 16,
  "qkv_bias": true,
  "transformers_version": "4.30.2"
}