hushem_40x_deit_small_sgd_0001_fold5
This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 1.1092
- Accuracy: 0.4634
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: 0.0001
- 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 |
|---|---|---|---|---|
| 1.7467 | 1.0 | 220 | 1.6098 | 0.2683 |
| 1.5306 | 2.0 | 440 | 1.5314 | 0.2683 |
| 1.3989 | 3.0 | 660 | 1.5004 | 0.2439 |
| 1.3588 | 4.0 | 880 | 1.4811 | 0.2195 |
| 1.3953 | 5.0 | 1100 | 1.4639 | 0.2683 |
| 1.3096 | 6.0 | 1320 | 1.4476 | 0.2439 |
| 1.2743 | 7.0 | 1540 | 1.4329 | 0.2683 |
| 1.2405 | 8.0 | 1760 | 1.4190 | 0.2927 |
| 1.253 | 9.0 | 1980 | 1.4052 | 0.3171 |
| 1.2253 | 10.0 | 2200 | 1.3912 | 0.3171 |
| 1.1663 | 11.0 | 2420 | 1.3767 | 0.3659 |
| 1.1699 | 12.0 | 2640 | 1.3616 | 0.3659 |
| 1.1615 | 13.0 | 2860 | 1.3463 | 0.3659 |
| 1.0999 | 14.0 | 3080 | 1.3303 | 0.3902 |
| 1.1286 | 15.0 | 3300 | 1.3148 | 0.3659 |
| 1.1333 | 16.0 | 3520 | 1.2990 | 0.3659 |
| 1.075 | 17.0 | 3740 | 1.2842 | 0.3659 |
| 1.0779 | 18.0 | 3960 | 1.2709 | 0.3659 |
| 1.0652 | 19.0 | 4180 | 1.2579 | 0.3659 |
| 1.0475 | 20.0 | 4400 | 1.2462 | 0.3659 |
| 1.0095 | 21.0 | 4620 | 1.2350 | 0.3902 |
| 1.0607 | 22.0 | 4840 | 1.2247 | 0.3902 |
| 1.0243 | 23.0 | 5060 | 1.2151 | 0.4146 |
| 1.0174 | 24.0 | 5280 | 1.2064 | 0.4146 |
| 0.9654 | 25.0 | 5500 | 1.1977 | 0.3902 |
| 1.017 | 26.0 | 5720 | 1.1899 | 0.4146 |
| 1.0002 | 27.0 | 5940 | 1.1820 | 0.3902 |
| 1.0191 | 28.0 | 6160 | 1.1750 | 0.3902 |
| 0.9876 | 29.0 | 6380 | 1.1683 | 0.3902 |
| 0.9526 | 30.0 | 6600 | 1.1623 | 0.4146 |
| 0.9957 | 31.0 | 6820 | 1.1566 | 0.4390 |
| 0.9778 | 32.0 | 7040 | 1.1513 | 0.4390 |
| 0.9223 | 33.0 | 7260 | 1.1464 | 0.4634 |
| 0.9281 | 34.0 | 7480 | 1.1418 | 0.4634 |
| 0.9107 | 35.0 | 7700 | 1.1376 | 0.4634 |
| 0.9485 | 36.0 | 7920 | 1.1336 | 0.4634 |
| 0.9035 | 37.0 | 8140 | 1.1298 | 0.4634 |
| 0.9223 | 38.0 | 8360 | 1.1266 | 0.4634 |
| 0.9312 | 39.0 | 8580 | 1.1235 | 0.4634 |
| 0.8782 | 40.0 | 8800 | 1.1209 | 0.4634 |
| 0.9252 | 41.0 | 9020 | 1.1184 | 0.4634 |
| 0.8989 | 42.0 | 9240 | 1.1164 | 0.4634 |
| 0.8959 | 43.0 | 9460 | 1.1145 | 0.4634 |
| 0.8589 | 44.0 | 9680 | 1.1130 | 0.4634 |
| 0.8899 | 45.0 | 9900 | 1.1117 | 0.4634 |
| 0.8915 | 46.0 | 10120 | 1.1107 | 0.4634 |
| 0.9043 | 47.0 | 10340 | 1.1100 | 0.4634 |
| 0.8309 | 48.0 | 10560 | 1.1095 | 0.4634 |
| 0.8724 | 49.0 | 10780 | 1.1093 | 0.4634 |
| 0.9011 | 50.0 | 11000 | 1.1092 | 0.4634 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.1.0+cu121
- Datasets 2.12.0
- Tokenizers 0.13.2