config.json

{
  "_name_or_path": "MCG-NJU/videomae-base",
  "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": "0_nihonbutou",
    "1": "10_nihonbutou",
    "2": "11_nihonbutou",
    "3": "12_nihonbutou",
    "4": "13_nihonbutou",
    "5": "14_nihonbutou",
    "6": "15_nihonbutou",
    "7": "16_nihonbutou",
    "8": "1_nihonbutou",
    "9": "2_nihonbutou",
    "10": "3_nihonbutou",
    "11": "4_nihonbutou",
    "12": "5_nihonbutou",
    "13": "6_nihonbutou",
    "14": "7_nihonbutou",
    "15": "8_nihonbutou",
    "16": "9_nihonbutou"
  },
  "image_size": 224,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "label2id": {
    "0_nihonbutou": 0,
    "10_nihonbutou": 1,
    "11_nihonbutou": 2,
    "12_nihonbutou": 3,
    "13_nihonbutou": 4,
    "14_nihonbutou": 5,
    "15_nihonbutou": 6,
    "16_nihonbutou": 7,
    "1_nihonbutou": 8,
    "2_nihonbutou": 9,
    "3_nihonbutou": 10,
    "4_nihonbutou": 11,
    "5_nihonbutou": 12,
    "6_nihonbutou": 13,
    "7_nihonbutou": 14,
    "8_nihonbutou": 15,
    "9_nihonbutou": 16
  },
  "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": "4.45.1",
  "tubelet_size": 2,
  "use_mean_pooling": false
}