{ "_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": "Abuse", "1": "Arrest", "2": "Arson", "3": "Assault", "4": "Burglary", "5": "Explosion", "6": "Fighting", "7": "Normal_Videos_for_Event_Recognition", "8": "RoadAccidents", "9": "Robbery", "10": "Shooting", "11": "Shoplifting", "12": "Stealing", "13": "Vandalism" }, "image_size": 224, "initializer_range": 0.02, "intermediate_size": 3072, "label2id": { "Abuse": 0, "Arrest": 1, "Arson": 2, "Assault": 3, "Burglary": 4, "Explosion": 5, "Fighting": 6, "Normal_Videos_for_Event_Recognition": 7, "RoadAccidents": 8, "Robbery": 9, "Shooting": 10, "Shoplifting": 11, "Stealing": 12, "Vandalism": 13 }, "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": "4.38.1", "tubelet_size": 2, "use_mean_pooling": false }