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": "teep",
    "1": "lead_kick",
    "2": "rear_knee",
    "3": "sidekick",
    "4": "cross",
    "5": "lead_hook",
    "6": "lead_elbow",
    "7": "rear_uppercut",
    "8": "rear_elbow",
    "9": "jab",
    "10": "rear_kick",
    "11": "check",
    "12": "lead_knee",
    "13": "lead_uppercut"
  },
  "image_size": 224,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "check": 11,
    "cross": 4,
    "jab": 9,
    "lead_elbow": 6,
    "lead_hook": 5,
    "lead_kick": 1,
    "lead_knee": 12,
    "lead_uppercut": 13,
    "rear_elbow": 8,
    "rear_kick": 10,
    "rear_knee": 2,
    "rear_uppercut": 7,
    "sidekick": 3,
    "teep": 0
  },
  "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
}