donut_synDB_test
This model is a fine-tuned version of naver-clova-ix/donut-base on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.1655
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: 2e-05
- train_batch_size: 2
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: inverse_sqrt
- lr_scheduler_warmup_steps: 480
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
0.3534 | 1.04 | 250 | 0.3525 |
0.265 | 1.15 | 275 | 0.3308 |
0.2096 | 1.25 | 300 | 0.2394 |
0.1623 | 1.35 | 325 | 0.2565 |
0.1336 | 1.46 | 350 | 0.1428 |
0.1114 | 1.56 | 375 | 0.2766 |
0.0873 | 1.67 | 400 | 0.2206 |
0.0688 | 1.77 | 425 | 0.2989 |
0.0769 | 1.88 | 450 | 0.2999 |
0.0947 | 1.98 | 475 | 0.2114 |
0.046 | 2.08 | 500 | 0.1818 |
0.051 | 2.19 | 525 | 0.2281 |
0.06 | 2.29 | 550 | 0.1434 |
0.0421 | 2.4 | 575 | 0.1299 |
0.0409 | 2.5 | 600 | 0.1819 |
0.0425 | 2.6 | 625 | 0.1042 |
0.0598 | 2.71 | 650 | 0.1145 |
0.034 | 2.81 | 675 | 0.1575 |
0.0268 | 2.92 | 700 | 0.1348 |
0.0509 | 3.02 | 725 | 0.1492 |
0.0205 | 3.12 | 750 | 0.0745 |
0.0381 | 3.23 | 775 | 0.1537 |
0.0227 | 3.33 | 800 | 0.0834 |
0.0161 | 3.44 | 825 | 0.1299 |
0.022 | 3.54 | 850 | 0.1248 |
0.0165 | 3.65 | 875 | 0.1059 |
0.019 | 3.75 | 900 | 0.1077 |
0.0179 | 3.85 | 925 | 0.1544 |
0.0237 | 3.96 | 950 | 0.1230 |
0.011 | 4.06 | 975 | 0.1396 |
0.0171 | 4.17 | 1000 | 0.1840 |
0.0108 | 4.27 | 1025 | 0.1513 |
0.0134 | 4.38 | 1050 | 0.2424 |
0.0289 | 4.48 | 1075 | 0.1194 |
0.0076 | 4.58 | 1100 | 0.1104 |
0.0086 | 4.69 | 1125 | 0.1079 |
0.0047 | 4.79 | 1150 | 0.1685 |
0.015 | 4.9 | 1175 | 0.1554 |
0.0173 | 5.0 | 1200 | 0.1315 |
0.0279 | 5.1 | 1225 | 0.1244 |
0.0111 | 5.21 | 1250 | 0.1655 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.2.2+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
- Downloads last month
- 109
Unable to determine this model’s pipeline type. Check the
docs
.