<!-- 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_ce_8
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: 0.8308
- Accuracy: 0.8822
- Brier Loss: 0.2169
- Nll: 0.9246
- F1 Micro: 0.8822
- F1 Macro: 0.8823
- Ece: 0.1044
- Aurc: 0.0265
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: 8
- eval_batch_size: 8
- 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 |
---|---|---|---|---|---|---|---|---|---|---|
0.2906 | 1.0 | 2000 | 0.5302 | 0.867 | 0.2106 | 1.2043 | 0.867 | 0.8685 | 0.0746 | 0.0284 |
0.236 | 2.0 | 4000 | 0.5819 | 0.8695 | 0.2142 | 1.1215 | 0.8695 | 0.8688 | 0.0909 | 0.0267 |
0.1236 | 3.0 | 6000 | 0.7115 | 0.8605 | 0.2390 | 1.1453 | 0.8605 | 0.8604 | 0.1069 | 0.0295 |
0.0703 | 4.0 | 8000 | 0.6965 | 0.8715 | 0.2265 | 1.0124 | 0.8715 | 0.8720 | 0.1015 | 0.0290 |
0.0307 | 5.0 | 10000 | 0.7503 | 0.8742 | 0.2229 | 0.9824 | 0.8742 | 0.8746 | 0.1052 | 0.0257 |
0.0229 | 6.0 | 12000 | 0.8042 | 0.874 | 0.2304 | 1.0125 | 0.874 | 0.8742 | 0.1091 | 0.0269 |
0.0114 | 7.0 | 14000 | 0.8335 | 0.8715 | 0.2283 | 1.0146 | 0.8715 | 0.8709 | 0.1103 | 0.0267 |
0.0082 | 8.0 | 16000 | 0.8655 | 0.873 | 0.2297 | 1.0222 | 0.8730 | 0.8735 | 0.1112 | 0.0279 |
0.002 | 9.0 | 18000 | 0.8350 | 0.8808 | 0.2180 | 0.9519 | 0.8808 | 0.8812 | 0.1067 | 0.0266 |
0.0041 | 10.0 | 20000 | 0.8308 | 0.8822 | 0.2169 | 0.9246 | 0.8822 | 0.8823 | 0.1044 | 0.0265 |
Framework versions
- Transformers 4.33.3
- Pytorch 2.2.0.dev20231002
- Datasets 2.7.1
- Tokenizers 0.13.3