config.json

{
  "_name_or_path": "MCG-NJU/videomae-base-finetuned-ssv2",
  "architectures": [
    "VideoMAEForVideoClassification"
  ],
  "attention_probs_dropout_prob": 0.0,
  "decoder_hidden_size": 384,
  "decoder_intermediate_size": 1536,
  "decoder_num_attention_heads": 6,
  "decoder_num_hidden_layers": 4,
  "hidden_act": "gelu",
  "hidden_dropout_prob": 0.0,
  "hidden_size": 768,
  "id2label": {
    "0": "badminton",
    "1": "clap",
    "2": "dance",
    "3": "drinking",
    "4": "football",
    "5": "handshake",
    "6": "handwave",
    "7": "jogging",
    "8": "jumping",
    "9": "pouring",
    "10": "pt_exercise",
    "11": "punching",
    "12": "pushups",
    "13": "rotating",
    "14": "running",
    "15": "situps",
    "16": "swirling",
    "17": "swirling_umbrella",
    "18": "volleyball",
    "19": "walking"
  },
  "image_size": 224,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "badminton": 0,
    "clap": 1,
    "dance": 2,
    "drinking": 3,
    "football": 4,
    "handshake": 5,
    "handwave": 6,
    "jogging": 7,
    "jumping": 8,
    "pouring": 9,
    "pt_exercise": 10,
    "punching": 11,
    "pushups": 12,
    "rotating": 13,
    "running": 14,
    "situps": 15,
    "swirling": 16,
    "swirling_umbrella": 17,
    "volleyball": 18,
    "walking": 19
  },
  "layer_norm_eps": 1e-12,
  "model_type": "videomae",
  "norm_pix_loss": true,
  "num_attention_heads": 12,
  "num_channels": 3,
  "num_frames": 16,
  "num_hidden_layers": 12,
  "patch_size": 16,
  "problem_type": "single_label_classification",
  "qkv_bias": true,
  "torch_dtype": "float32",
  "transformers_version": "4.33.2",
  "tubelet_size": 2,
  "use_mean_pooling": true
}