FFPP-Raw_1FPS_faces-expand-0-aligned
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: 0.9932
- Precision: 0.9994
- 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 |
---|---|---|---|---|---|---|---|---|
0.0983 | 1.0 | 1377 | 0.0679 | 0.9743 | 0.9700 | 0.9165 | 0.9425 | 0.9961 |
0.0917 | 2.0 | 2755 | 0.0342 | 0.9896 | 0.9718 | 0.9803 | 0.9760 | 0.9993 |
0.0291 | 3.0 | 4132 | 0.0161 | 0.9940 | 0.9908 | 0.9818 | 0.9863 | 0.9998 |
0.0454 | 4.0 | 5510 | 0.0136 | 0.9950 | 0.9851 | 0.9917 | 0.9884 | 0.9998 |
0.0302 | 5.0 | 6887 | 0.0075 | 0.9972 | 0.9896 | 0.9976 | 0.9936 | 1.0000 |
0.0073 | 6.0 | 8265 | 0.0064 | 0.9976 | 0.9931 | 0.9957 | 0.9944 | 1.0000 |
0.016 | 7.0 | 9642 | 0.0067 | 0.9975 | 0.9934 | 0.9949 | 0.9941 | 1.0000 |
0.0054 | 8.0 | 11020 | 0.0058 | 0.9978 | 0.9915 | 0.9984 | 0.9949 | 1.0000 |
0.0237 | 9.0 | 12397 | 0.0063 | 0.9975 | 0.9894 | 0.9993 | 0.9943 | 1.0000 |
0.0088 | 10.0 | 13775 | 0.0042 | 0.9982 | 0.9920 | 0.9995 | 0.9957 | 1.0000 |
0.0078 | 11.0 | 15152 | 0.0043 | 0.9982 | 0.9921 | 0.9994 | 0.9957 | 1.0000 |
0.0142 | 12.0 | 16530 | 0.0040 | 0.9982 | 0.9939 | 0.9979 | 0.9959 | 1.0000 |
0.0058 | 13.0 | 17907 | 0.0035 | 0.9983 | 0.9930 | 0.9992 | 0.9961 | 1.0000 |
0.0076 | 14.0 | 19285 | 0.0040 | 0.9981 | 0.9920 | 0.9994 | 0.9957 | 1.0000 |
0.0032 | 15.0 | 20662 | 0.0036 | 0.9983 | 0.9926 | 0.9995 | 0.9960 | 1.0000 |
0.0154 | 16.0 | 22040 | 0.0033 | 0.9983 | 0.9928 | 0.9996 | 0.9962 | 1.0000 |
0.0041 | 17.0 | 23417 | 0.0032 | 0.9984 | 0.9925 | 0.9999 | 0.9962 | 1.0000 |
0.002 | 18.0 | 24795 | 0.0032 | 0.9984 | 0.9933 | 0.9992 | 0.9962 | 1.0000 |
0.0024 | 19.0 | 26172 | 0.0031 | 0.9984 | 0.9932 | 0.9994 | 0.9963 | 1.0000 |
0.0023 | 19.99 | 27540 | 0.0031 | 0.9984 | 0.9927 | 0.9998 | 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
- 23
Finetuned from
Evaluation results
- Accuracy on imagefoldertest set self-reported0.998
- Recall on imagefoldertest set self-reported0.993
- Precision on imagefoldertest set self-reported0.999
- F1 on imagefoldertest set self-reported0.996