hushem_40x_deit_tiny_sgd_001_fold1
This model is a fine-tuned version of facebook/deit-tiny-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.9268
- Accuracy: 0.7333
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: 0.001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 50
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
1.1534 | 1.0 | 215 | 1.3121 | 0.3778 |
0.8744 | 2.0 | 430 | 1.2305 | 0.5111 |
0.7578 | 3.0 | 645 | 1.1023 | 0.5556 |
0.6354 | 4.0 | 860 | 0.9584 | 0.5778 |
0.4832 | 5.0 | 1075 | 0.8877 | 0.6444 |
0.4506 | 6.0 | 1290 | 0.8214 | 0.6889 |
0.3619 | 7.0 | 1505 | 0.8077 | 0.6889 |
0.3187 | 8.0 | 1720 | 0.7845 | 0.6667 |
0.2423 | 9.0 | 1935 | 0.7629 | 0.7111 |
0.2351 | 10.0 | 2150 | 0.7464 | 0.7333 |
0.2043 | 11.0 | 2365 | 0.7249 | 0.6889 |
0.1712 | 12.0 | 2580 | 0.7297 | 0.7111 |
0.1294 | 13.0 | 2795 | 0.7280 | 0.7333 |
0.1185 | 14.0 | 3010 | 0.7610 | 0.7333 |
0.1264 | 15.0 | 3225 | 0.7479 | 0.7333 |
0.0869 | 16.0 | 3440 | 0.7617 | 0.7333 |
0.0902 | 17.0 | 3655 | 0.7623 | 0.7333 |
0.0782 | 18.0 | 3870 | 0.7805 | 0.7333 |
0.071 | 19.0 | 4085 | 0.7715 | 0.7333 |
0.063 | 20.0 | 4300 | 0.7777 | 0.7333 |
0.0587 | 21.0 | 4515 | 0.7497 | 0.7333 |
0.0675 | 22.0 | 4730 | 0.7998 | 0.7333 |
0.0426 | 23.0 | 4945 | 0.8200 | 0.7333 |
0.0373 | 24.0 | 5160 | 0.8281 | 0.7111 |
0.0441 | 25.0 | 5375 | 0.8317 | 0.7111 |
0.0323 | 26.0 | 5590 | 0.8133 | 0.7111 |
0.0359 | 27.0 | 5805 | 0.8214 | 0.7111 |
0.0291 | 28.0 | 6020 | 0.8265 | 0.7111 |
0.0287 | 29.0 | 6235 | 0.8490 | 0.7111 |
0.0271 | 30.0 | 6450 | 0.8534 | 0.7111 |
0.0256 | 31.0 | 6665 | 0.8626 | 0.7111 |
0.0212 | 32.0 | 6880 | 0.8791 | 0.7111 |
0.0155 | 33.0 | 7095 | 0.8740 | 0.7333 |
0.0144 | 34.0 | 7310 | 0.8433 | 0.7333 |
0.0132 | 35.0 | 7525 | 0.8680 | 0.7333 |
0.015 | 36.0 | 7740 | 0.8880 | 0.7333 |
0.0129 | 37.0 | 7955 | 0.8931 | 0.7333 |
0.018 | 38.0 | 8170 | 0.8891 | 0.7333 |
0.0092 | 39.0 | 8385 | 0.9122 | 0.7333 |
0.0085 | 40.0 | 8600 | 0.9159 | 0.7333 |
0.0124 | 41.0 | 8815 | 0.9199 | 0.7333 |
0.0125 | 42.0 | 9030 | 0.9056 | 0.7333 |
0.0107 | 43.0 | 9245 | 0.9191 | 0.7333 |
0.0095 | 44.0 | 9460 | 0.9083 | 0.7333 |
0.0115 | 45.0 | 9675 | 0.9189 | 0.7333 |
0.0088 | 46.0 | 9890 | 0.9241 | 0.7333 |
0.0065 | 47.0 | 10105 | 0.9299 | 0.7333 |
0.007 | 48.0 | 10320 | 0.9257 | 0.7333 |
0.0129 | 49.0 | 10535 | 0.9260 | 0.7333 |
0.0229 | 50.0 | 10750 | 0.9268 | 0.7333 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.1+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2
- Downloads last month
- 6