videomae-base-finetuned-kinetics-0402_final_253259_roi

This model is a fine-tuned version of MCG-NJU/videomae-base-finetuned-kinetics on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.8553
• Accuracy: 0.7729

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• training_steps: 80100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.0115	0.0100	802	0.7706	0.8360
0.0058	1.0100	1604	0.8429	0.8177
0.1527	2.0100	2406	0.3902	0.8818
0.0121	3.0100	3208	0.5051	0.8673
0.3265	4.0100	4010	0.7498	0.7941
0.0034	5.0100	4812	0.5008	0.8581
0.0327	6.0100	5614	0.5419	0.8513
0.4163	7.0100	6416	0.8579	0.7879
0.2185	8.0100	7218	0.6873	0.7910
0.0013	9.0100	8020	1.2436	0.7605
0.0051	10.0100	8822	1.4545	0.7658
0.0334	11.0100	9624	1.2327	0.7071
0.0013	12.0100	10426	0.7517	0.8261
0.0048	13.0100	11228	0.6028	0.8497
1.5951	14.0100	12030	0.8834	0.7582
0.1437	15.0100	12832	1.3627	0.7361
0.0634	16.0100	13634	1.0872	0.7757
0.5734	17.0100	14436	0.9348	0.7757

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0