FFPP-Raw_1FPS_faces-expand-0-aligned-normalize-image-mean-std
This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.0031
- Accuracy: 0.9984
- Recall: 0.9926
- Precision: 0.9999
- F1: 0.9963
- Roc Auc: 1.0000
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
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 20
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Recall | Precision | F1 | Roc Auc |
---|---|---|---|---|---|---|---|---|
0.1087 | 1.0 | 1377 | 0.0767 | 0.9720 | 0.9020 | 0.9667 | 0.9332 | 0.9965 |
0.0647 | 2.0 | 2755 | 0.0408 | 0.9847 | 0.9458 | 0.9831 | 0.9641 | 0.9990 |
0.0295 | 3.0 | 4132 | 0.0499 | 0.9862 | 0.9947 | 0.9449 | 0.9692 | 0.9995 |
0.0537 | 4.0 | 5510 | 0.0201 | 0.9927 | 0.9736 | 0.9928 | 0.9831 | 0.9998 |
0.0253 | 5.0 | 6887 | 0.0115 | 0.9958 | 0.9826 | 0.9979 | 0.9902 | 0.9999 |
0.0069 | 6.0 | 8265 | 0.0058 | 0.9979 | 0.9928 | 0.9974 | 0.9951 | 1.0000 |
0.0226 | 7.0 | 9642 | 0.0113 | 0.9960 | 0.9823 | 0.9994 | 0.9908 | 0.9999 |
0.0096 | 8.0 | 11020 | 0.0147 | 0.9957 | 0.9806 | 0.9995 | 0.9900 | 0.9999 |
0.0301 | 9.0 | 12397 | 0.0071 | 0.9972 | 0.9878 | 0.9993 | 0.9935 | 1.0000 |
0.0133 | 10.0 | 13775 | 0.0055 | 0.9978 | 0.9901 | 0.9998 | 0.9949 | 1.0000 |
0.0074 | 11.0 | 15152 | 0.0049 | 0.9980 | 0.9915 | 0.9992 | 0.9953 | 1.0000 |
0.0036 | 12.0 | 16530 | 0.0034 | 0.9983 | 0.9935 | 0.9987 | 0.9961 | 1.0000 |
0.0039 | 13.0 | 17907 | 0.0037 | 0.9982 | 0.9926 | 0.9993 | 0.9959 | 1.0000 |
0.0106 | 14.0 | 19285 | 0.0033 | 0.9983 | 0.9931 | 0.9992 | 0.9961 | 1.0000 |
0.0033 | 15.0 | 20662 | 0.0033 | 0.9984 | 0.9926 | 0.9998 | 0.9962 | 1.0000 |
0.0214 | 16.0 | 22040 | 0.0032 | 0.9984 | 0.9931 | 0.9994 | 0.9963 | 1.0000 |
0.0041 | 17.0 | 23417 | 0.0032 | 0.9984 | 0.9926 | 0.9998 | 0.9962 | 1.0000 |
0.0101 | 18.0 | 24795 | 0.0031 | 0.9984 | 0.9926 | 0.9999 | 0.9963 | 1.0000 |
0.0023 | 19.0 | 26172 | 0.0031 | 0.9984 | 0.9928 | 0.9997 | 0.9963 | 1.0000 |
0.002 | 19.99 | 27540 | 0.0031 | 0.9984 | 0.9926 | 0.9999 | 0.9963 | 1.0000 |
Framework versions
- Transformers 4.39.2
- Pytorch 2.2.2
- Datasets 2.18.0
- Tokenizers 0.15.2
- Downloads last month
- 16
Finetuned from
Evaluation results
- Accuracy on imagefoldertest set self-reported0.998
- Recall on imagefoldertest set self-reported0.993
- Precision on imagefoldertest set self-reported1.000
- F1 on imagefoldertest set self-reported0.996