videomae-base-finetuned-ASBD_v3
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: 1.1126
- Accuracy: 0.6471
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: 12
- eval_batch_size: 12
- 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: 500
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.3497 | 0.1 | 49 | 1.6995 | 0.1757 |
| 1.2902 | 1.1 | 98 | 1.4059 | 0.2568 |
| 0.7943 | 2.1 | 147 | 1.1542 | 0.5135 |
| 0.775 | 3.1 | 196 | 1.2426 | 0.5946 |
| 0.4444 | 4.1 | 245 | 2.0844 | 0.4595 |
| 0.3258 | 5.1 | 294 | 1.2767 | 0.6351 |
| 0.1751 | 6.1 | 343 | 1.9364 | 0.5676 |
| 0.029 | 7.1 | 392 | 1.7154 | 0.5811 |
| 0.0115 | 8.1 | 441 | 1.6968 | 0.6622 |
| 0.023 | 9.1 | 490 | 1.6859 | 0.6486 |
| 0.0072 | 10.02 | 500 | 1.6885 | 0.6486 |
Framework versions
- Transformers 4.28.0
- Pytorch 2.1.0+cu118
- Datasets 2.14.6
- Tokenizers 0.13.3