config.json

{
  "_name_or_path": "MCG-NJU/videomae-base-finetuned-kinetics",
  "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": "SK",
    "1": "lupper",
    "2": "LBRound",
    "3": "rhook",
    "4": "LElbow",
    "5": "Teep",
    "6": "rupper",
    "7": "cross",
    "8": "HM",
    "9": "LKnee",
    "10": "lhook",
    "11": "jab",
    "12": "RKnee",
    "13": "Check",
    "14": "RLRound",
    "15": "RElbow",
    "16": "RBRound",
    "17": "Sidekick"
  },
  "image_size": 224,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "Check": 13,
    "HM": 8,
    "LBRound": 2,
    "LElbow": 4,
    "LKnee": 9,
    "RBRound": 16,
    "RElbow": 15,
    "RKnee": 12,
    "RLRound": 14,
    "SK": 0,
    "Sidekick": 17,
    "Teep": 5,
    "cross": 7,
    "jab": 11,
    "lhook": 10,
    "lupper": 1,
    "rhook": 3,
    "rupper": 6
  },
  "layer_norm_eps": 1e-12,
  "model_type": "videomae",
  "norm_pix_loss": false,
  "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": "4.47.0.dev0",
  "tubelet_size": 2,
  "use_mean_pooling": true
}