videomae-base-finetuned-kinetics-finetuned-data-no-yolo-kaggle
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: 0.4866
- Accuracy: 0.9046
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: 9e-05
- train_batch_size: 2
- eval_batch_size: 2
- 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: 4125
Training results
Training Loss |
Epoch |
Step |
Validation Loss |
Accuracy |
0.2716 |
0.09 |
376 |
0.6657 |
0.8147 |
0.2248 |
1.09 |
752 |
0.5673 |
0.8065 |
0.0106 |
2.09 |
1128 |
0.4900 |
0.8529 |
0.0548 |
3.09 |
1504 |
0.5760 |
0.8719 |
0.0203 |
4.09 |
1880 |
0.8674 |
0.8311 |
0.0008 |
5.09 |
2256 |
0.5422 |
0.8692 |
0.0004 |
6.09 |
2632 |
0.4938 |
0.8965 |
0.0021 |
7.09 |
3008 |
0.7053 |
0.8583 |
0.0002 |
8.09 |
3384 |
0.5675 |
0.8747 |
0.0003 |
9.09 |
3760 |
0.4930 |
0.9046 |
0.0003 |
10.09 |
4125 |
0.4866 |
0.9046 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.2.2+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2