config.json

{
  "_name_or_path": "google/vivit-b-16x2-kinetics400",
  "architectures": [
    "VivitForVideoClassification"
  ],
  "attention_probs_dropout_prob": 0.0,
  "hidden_act": "gelu_fast",
  "hidden_dropout_prob": 0.0,
  "hidden_size": 768,
  "id2label": {
    "0": "Abuse",
    "1": "Arrest",
    "2": "Arson",
    "3": "Assault",
    "4": "Burglary",
    "5": "Explosion",
    "6": "Fighting",
    "7": "RoadAccidents",
    "8": "Robbery",
    "9": "Shooting",
    "10": "Shoplifting",
    "11": "Stealing",
    "12": "Vandalism"
  },
  "image_size": 224,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "Abuse": 0,
    "Arrest": 1,
    "Arson": 2,
    "Assault": 3,
    "Burglary": 4,
    "Explosion": 5,
    "Fighting": 6,
    "RoadAccidents": 7,
    "Robbery": 8,
    "Shooting": 9,
    "Shoplifting": 10,
    "Stealing": 11,
    "Vandalism": 12
  },
  "layer_norm_eps": 1e-06,
  "model_type": "vivit",
  "num_attention_heads": 12,
  "num_channels": 3,
  "num_frames": 32,
  "num_hidden_layers": 12,
  "qkv_bias": true,
  "torch_dtype": "float32",
  "transformers_version": "4.46.2",
  "tubelet_size": [
    2,
    16,
    16
  ],
  "video_size": [
    32,
    224,
    224
  ]
}