sewd-classifier-aug
This model is a fine-tuned version of asapp/sew-d-tiny-100k-ft-ls100h on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.3679
- Accuracy: 0.6092
- Precision: 0.6385
- Recall: 0.6092
- F1: 0.5780
- Binary: 0.7259
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: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 128
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 | Binary |
---|---|---|---|---|---|---|---|---|
No log | 0.19 | 50 | 4.3617 | 0.0431 | 0.0054 | 0.0431 | 0.0091 | 0.2267 |
No log | 0.38 | 100 | 4.1063 | 0.0593 | 0.0134 | 0.0593 | 0.0187 | 0.3046 |
No log | 0.58 | 150 | 3.8795 | 0.1105 | 0.0541 | 0.1105 | 0.0502 | 0.3553 |
No log | 0.77 | 200 | 3.6851 | 0.1294 | 0.0368 | 0.1294 | 0.0511 | 0.3768 |
No log | 0.96 | 250 | 3.5025 | 0.2022 | 0.1342 | 0.2022 | 0.1289 | 0.4337 |
4.0829 | 1.15 | 300 | 3.3027 | 0.2049 | 0.1255 | 0.2049 | 0.1208 | 0.4394 |
4.0829 | 1.34 | 350 | 3.1674 | 0.2291 | 0.1067 | 0.2291 | 0.1305 | 0.4571 |
4.0829 | 1.53 | 400 | 3.0183 | 0.2453 | 0.1576 | 0.2453 | 0.1515 | 0.4677 |
4.0829 | 1.73 | 450 | 2.9047 | 0.2830 | 0.1807 | 0.2830 | 0.1898 | 0.4949 |
4.0829 | 1.92 | 500 | 2.7836 | 0.3181 | 0.2598 | 0.3181 | 0.2400 | 0.5194 |
3.3139 | 2.11 | 550 | 2.6784 | 0.3396 | 0.2432 | 0.3396 | 0.2484 | 0.5345 |
3.3139 | 2.3 | 600 | 2.5843 | 0.3261 | 0.2363 | 0.3261 | 0.2332 | 0.5259 |
3.3139 | 2.49 | 650 | 2.5050 | 0.3288 | 0.2625 | 0.3288 | 0.2465 | 0.5286 |
3.3139 | 2.68 | 700 | 2.3782 | 0.3531 | 0.3019 | 0.3531 | 0.2844 | 0.5456 |
3.3139 | 2.88 | 750 | 2.2826 | 0.4124 | 0.3926 | 0.4124 | 0.3473 | 0.5871 |
2.8692 | 3.07 | 800 | 2.2188 | 0.4151 | 0.3627 | 0.4151 | 0.3390 | 0.5889 |
2.8692 | 3.26 | 850 | 2.1541 | 0.4124 | 0.3370 | 0.4124 | 0.3348 | 0.5871 |
2.8692 | 3.45 | 900 | 2.0925 | 0.4016 | 0.3462 | 0.4016 | 0.3324 | 0.5787 |
2.8692 | 3.64 | 950 | 2.0181 | 0.4286 | 0.3550 | 0.4286 | 0.3612 | 0.5984 |
2.8692 | 3.84 | 1000 | 1.9575 | 0.4663 | 0.4514 | 0.4663 | 0.4074 | 0.6248 |
2.5712 | 4.03 | 1050 | 1.9088 | 0.4771 | 0.4544 | 0.4771 | 0.4229 | 0.6323 |
2.5712 | 4.22 | 1100 | 1.8500 | 0.4906 | 0.4284 | 0.4906 | 0.4235 | 0.6418 |
2.5712 | 4.41 | 1150 | 1.8270 | 0.4852 | 0.4312 | 0.4852 | 0.4222 | 0.6380 |
2.5712 | 4.6 | 1200 | 1.7758 | 0.4852 | 0.4556 | 0.4852 | 0.4306 | 0.6380 |
2.5712 | 4.79 | 1250 | 1.7430 | 0.4933 | 0.4552 | 0.4933 | 0.4388 | 0.6437 |
2.5712 | 4.99 | 1300 | 1.7188 | 0.5067 | 0.4941 | 0.5067 | 0.4624 | 0.6531 |
2.3754 | 5.18 | 1350 | 1.6813 | 0.5148 | 0.4820 | 0.5148 | 0.4668 | 0.6580 |
2.3754 | 5.37 | 1400 | 1.6463 | 0.5472 | 0.5302 | 0.5472 | 0.5029 | 0.6806 |
2.3754 | 5.56 | 1450 | 1.6446 | 0.5256 | 0.5247 | 0.5256 | 0.4801 | 0.6655 |
2.3754 | 5.75 | 1500 | 1.6005 | 0.5660 | 0.5579 | 0.5660 | 0.5235 | 0.6930 |
2.3754 | 5.94 | 1550 | 1.5667 | 0.5795 | 0.5561 | 0.5795 | 0.5363 | 0.7032 |
2.234 | 6.14 | 1600 | 1.5397 | 0.5741 | 0.5389 | 0.5741 | 0.5291 | 0.6995 |
2.234 | 6.33 | 1650 | 1.5235 | 0.5687 | 0.5444 | 0.5687 | 0.5219 | 0.6957 |
2.234 | 6.52 | 1700 | 1.5045 | 0.5849 | 0.5699 | 0.5849 | 0.5376 | 0.7070 |
2.234 | 6.71 | 1750 | 1.4936 | 0.5822 | 0.5635 | 0.5822 | 0.5383 | 0.7059 |
2.234 | 6.9 | 1800 | 1.4688 | 0.5822 | 0.5736 | 0.5822 | 0.5401 | 0.7059 |
2.1257 | 7.09 | 1850 | 1.4533 | 0.5957 | 0.6042 | 0.5957 | 0.5536 | 0.7154 |
2.1257 | 7.29 | 1900 | 1.4233 | 0.6038 | 0.6097 | 0.6038 | 0.5675 | 0.7202 |
2.1257 | 7.48 | 1950 | 1.4267 | 0.6038 | 0.6224 | 0.6038 | 0.5721 | 0.7210 |
2.1257 | 7.67 | 2000 | 1.4075 | 0.6146 | 0.6310 | 0.6146 | 0.5804 | 0.7278 |
2.1257 | 7.86 | 2050 | 1.3887 | 0.5984 | 0.6178 | 0.5984 | 0.5610 | 0.7173 |
2.0636 | 8.05 | 2100 | 1.3679 | 0.6092 | 0.6385 | 0.6092 | 0.5780 | 0.7259 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.3.0
- Datasets 2.19.1
- Tokenizers 0.15.1
- Downloads last month
- 1