{
  "_name_or_path": "google/vit-base-patch16-224",
  "architectures": [
    "ViTForImageClassification"
  ],
  "attention_probs_dropout_prob": 0.0,
  "encoder_stride": 16,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.0,
  "hidden_size": 768,
  "id2label": {
    "0": "aeroplane",
    "1": "blackboard",
    "10": "clouds",
    "11": "coins",
    "12": "cycles",
    "13": "deer",
    "14": "desk",
    "15": "dogs",
    "16": "dogsledge",
    "17": "door-frame",
    "18": "factory",
    "19": "fireman",
    "2": "boat",
    "20": "firetruck",
    "21": "food",
    "22": "helicopter",
    "23": "horse",
    "24": "horsepipe",
    "25": "instrument",
    "26": "ladder",
    "27": "lake",
    "28": "landscape",
    "29": "machinaries",
    "3": "books",
    "30": "mountains",
    "31": "painting",
    "32": "people",
    "33": "pole",
    "34": "railway",
    "35": "river",
    "36": "road",
    "37": "ship",
    "38": "ski",
    "39": "sky",
    "4": "bridge",
    "40": "snow",
    "41": "stairs",
    "42": "table",
    "43": "telephone",
    "44": "tree",
    "45": "water",
    "5": "building",
    "6": "car",
    "7": "chair",
    "8": "children",
    "9": "church"
  },
  "image_size": 224,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "aeroplane": "0",
    "blackboard": "1",
    "boat": "2",
    "books": "3",
    "bridge": "4",
    "building": "5",
    "car": "6",
    "chair": "7",
    "children": "8",
    "church": "9",
    "clouds": "10",
    "coins": "11",
    "cycles": "12",
    "deer": "13",
    "desk": "14",
    "dogs": "15",
    "dogsledge": "16",
    "door-frame": "17",
    "factory": "18",
    "fireman": "19",
    "firetruck": "20",
    "food": "21",
    "helicopter": "22",
    "horse": "23",
    "horsepipe": "24",
    "instrument": "25",
    "ladder": "26",
    "lake": "27",
    "landscape": "28",
    "machinaries": "29",
    "mountains": "30",
    "painting": "31",
    "people": "32",
    "pole": "33",
    "railway": "34",
    "river": "35",
    "road": "36",
    "ship": "37",
    "ski": "38",
    "sky": "39",
    "snow": "40",
    "stairs": "41",
    "table": "42",
    "telephone": "43",
    "tree": "44",
    "water": "45"
  },
  "layer_norm_eps": 1e-12,
  "model_type": "vit",
  "num_attention_heads": 12,
  "num_channels": 3,
  "num_hidden_layers": 12,
  "patch_size": 16,
  "problem_type": "single_label_classification",
  "qkv_bias": true,
  "torch_dtype": "float32",
  "transformers_version": "4.37.2"
}