config.json

{
  "_name_or_path": "ak3ra/videomae-base-finetuned-kinetics-finetuned-ntu-model",
  "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": "back_pain",
    "1": "chest_pain",
    "2": "drink_water",
    "3": "eat_meal",
    "4": "falling_down",
    "5": "hand_waving",
    "6": "headache",
    "7": "nausea_vomiting",
    "8": "neck_pain",
    "9": "play_with_phone_tablet",
    "10": "put_on_jacket",
    "11": "reading",
    "12": "sit_down",
    "13": "sneeze_cough",
    "14": "staggering",
    "15": "stand_up",
    "16": "take_off_jacket"
  },
  "image_size": 224,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "back_pain": 0,
    "chest_pain": 1,
    "drink_water": 2,
    "eat_meal": 3,
    "falling_down": 4,
    "hand_waving": 5,
    "headache": 6,
    "nausea_vomiting": 7,
    "neck_pain": 8,
    "play_with_phone_tablet": 9,
    "put_on_jacket": 10,
    "reading": 11,
    "sit_down": 12,
    "sneeze_cough": 13,
    "staggering": 14,
    "stand_up": 15,
    "take_off_jacket": 16
  },
  "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": "4.27.0.dev0",
  "tubelet_size": 2,
  "use_mean_pooling": true
}