<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
vit-base_rvl_cdip-N1K_aAURC_16
This model is a fine-tuned version of jordyvl/vit-base_rvl-cdip on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.5629
- Accuracy: 0.8892
- Brier Loss: 0.1995
- Nll: 0.8643
- F1 Micro: 0.8892
- F1 Macro: 0.8898
- Ece: 0.0923
- Aurc: 0.0215
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: 16
- eval_batch_size: 16
- 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: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Brier Loss | Nll | F1 Micro | F1 Macro | Ece | Aurc |
|---|---|---|---|---|---|---|---|---|---|---|
| 0.1794 | 1.0 | 1000 | 0.3827 | 0.8815 | 0.1829 | 1.1942 | 0.8815 | 0.8822 | 0.0573 | 0.0226 |
| 0.1415 | 2.0 | 2000 | 0.4705 | 0.8698 | 0.2118 | 1.1615 | 0.8698 | 0.8686 | 0.0859 | 0.0259 |
| 0.0725 | 3.0 | 3000 | 0.4582 | 0.8768 | 0.1996 | 1.0476 | 0.8768 | 0.8771 | 0.0845 | 0.0234 |
| 0.0388 | 4.0 | 4000 | 0.4958 | 0.879 | 0.2024 | 1.0000 | 0.879 | 0.8798 | 0.0877 | 0.0259 |
| 0.0153 | 5.0 | 5000 | 0.5171 | 0.8815 | 0.2047 | 0.9580 | 0.8815 | 0.8815 | 0.0942 | 0.0229 |
| 0.0069 | 6.0 | 6000 | 0.5334 | 0.8845 | 0.2021 | 0.9350 | 0.8845 | 0.8854 | 0.0922 | 0.0230 |
| 0.005 | 7.0 | 7000 | 0.5412 | 0.8905 | 0.1964 | 0.9179 | 0.8905 | 0.8907 | 0.0886 | 0.0218 |
| 0.0043 | 8.0 | 8000 | 0.5497 | 0.8892 | 0.1985 | 0.8970 | 0.8892 | 0.8900 | 0.0901 | 0.0225 |
| 0.0023 | 9.0 | 9000 | 0.5610 | 0.8878 | 0.1994 | 0.8679 | 0.8878 | 0.8883 | 0.0932 | 0.0220 |
| 0.0024 | 10.0 | 10000 | 0.5629 | 0.8892 | 0.1995 | 0.8643 | 0.8892 | 0.8898 | 0.0923 | 0.0215 |
Framework versions
- Transformers 4.33.3
- Pytorch 2.2.0.dev20231002
- Datasets 2.7.1
- Tokenizers 0.13.3
