metadata
license: mit
tags:
- generated_from_trainer
datasets:
- imagefolder
model-index:
- name: donut_pdf_ocr
results: []
donut_pdf_ocr
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: 0.0821
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: 2
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
0.6648 | 1.0 | 47 | 0.2550 |
0.3772 | 2.0 | 94 | 0.0895 |
0.1732 | 3.0 | 141 | 0.0862 |
0.1244 | 4.0 | 188 | 0.0718 |
0.0867 | 5.0 | 235 | 0.0473 |
0.1035 | 6.0 | 282 | 0.0640 |
0.1703 | 7.0 | 329 | 0.0877 |
0.0324 | 8.0 | 376 | 0.0722 |
0.033 | 9.0 | 423 | 0.0701 |
0.0404 | 10.0 | 470 | 0.0627 |
0.0198 | 11.0 | 517 | 0.0770 |
0.007 | 12.0 | 564 | 0.0730 |
0.0207 | 13.0 | 611 | 0.0885 |
0.0528 | 14.0 | 658 | 0.0898 |
0.0189 | 15.0 | 705 | 0.0588 |
0.0977 | 16.0 | 752 | 0.0563 |
0.0542 | 17.0 | 799 | 0.0647 |
0.0114 | 18.0 | 846 | 0.0674 |
0.0056 | 19.0 | 893 | 0.0726 |
0.005 | 20.0 | 940 | 0.0797 |
0.0005 | 21.0 | 987 | 0.0704 |
0.0023 | 22.0 | 1034 | 0.0864 |
0.0105 | 23.0 | 1081 | 0.0830 |
0.0014 | 24.0 | 1128 | 0.0806 |
0.0006 | 25.0 | 1175 | 0.0826 |
0.0012 | 26.0 | 1222 | 0.0790 |
0.0006 | 27.0 | 1269 | 0.0751 |
0.0022 | 28.0 | 1316 | 0.0786 |
0.0075 | 29.0 | 1363 | 0.0816 |
0.0024 | 30.0 | 1410 | 0.0821 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3