videomae-base-finetuned-ucf101-subset
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: 0.4364
- Accuracy: 0.8839
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- training_steps: 148
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 2.3351 | 0.13 | 19 | 2.0899 | 0.3286 |
| 1.9054 | 1.13 | 38 | 1.4402 | 0.5286 |
| 0.9695 | 2.13 | 57 | 0.7842 | 0.7143 |
| 0.4545 | 3.13 | 76 | 0.4825 | 0.8 |
| 0.3444 | 4.13 | 95 | 0.3753 | 0.8714 |
| 0.1893 | 5.13 | 114 | 0.3679 | 0.8429 |
| 0.1271 | 6.13 | 133 | 0.2520 | 0.9143 |
| 0.0856 | 7.1 | 148 | 0.2427 | 0.9143 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.1.0+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2