videomae-base-finetuned-judo

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

• Loss: 0.9544
• Accuracy: 0.6216

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use OptimizerNames.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: 0.1
• training_steps: 780

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.1454	0.1013	79	1.1138	0.4324
0.9601	1.1013	158	1.2666	0.3514
0.7864	2.1013	237	1.1963	0.4054
1.0221	3.1013	316	1.1635	0.3514
0.8762	4.1013	395	1.0229	0.4595
0.9761	5.1013	474	0.9012	0.5676
0.6493	6.1013	553	1.0638	0.6486
0.7152	7.1013	632	1.0122	0.6216
0.3646	8.1013	711	0.8835	0.6216
0.2113	9.0885	780	0.9544	0.6216

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0