<!-- 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_crl_softmax_rank1_fixed
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: 0.7091
- Accuracy: 0.9032
- Brier Loss: 0.1756
- Nll: 1.0964
- F1 Micro: 0.9032
- F1 Macro: 0.9033
- Ece: 0.0854
- Aurc: 0.0188
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
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- 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 |
---|---|---|---|---|---|---|---|---|---|---|
0.2289 | 1.0 | 10000 | 0.4298 | 0.8826 | 0.1794 | 1.2461 | 0.8826 | 0.8841 | 0.0553 | 0.0199 |
0.1972 | 2.0 | 20000 | 0.4350 | 0.8859 | 0.1769 | 1.3140 | 0.8859 | 0.8862 | 0.0558 | 0.0197 |
0.1414 | 3.0 | 30000 | 0.4423 | 0.8938 | 0.1702 | 1.2433 | 0.8938 | 0.8948 | 0.0639 | 0.0181 |
0.0903 | 4.0 | 40000 | 0.5076 | 0.8943 | 0.1753 | 1.2033 | 0.8943 | 0.8941 | 0.0766 | 0.0181 |
0.0684 | 5.0 | 50000 | 0.5592 | 0.8963 | 0.1783 | 1.2422 | 0.8963 | 0.8965 | 0.0811 | 0.0194 |
0.0313 | 6.0 | 60000 | 0.6384 | 0.8956 | 0.1836 | 1.2359 | 0.8957 | 0.8957 | 0.0861 | 0.0218 |
0.0163 | 7.0 | 70000 | 0.6673 | 0.9005 | 0.1788 | 1.1927 | 0.9005 | 0.9006 | 0.0855 | 0.0215 |
0.0104 | 8.0 | 80000 | 0.6929 | 0.9001 | 0.1791 | 1.1768 | 0.9001 | 0.9000 | 0.0860 | 0.0204 |
0.0036 | 9.0 | 90000 | 0.7131 | 0.9018 | 0.1780 | 1.1295 | 0.9018 | 0.9018 | 0.0866 | 0.0195 |
0.0023 | 10.0 | 100000 | 0.7091 | 0.9032 | 0.1756 | 1.0964 | 0.9032 | 0.9033 | 0.0854 | 0.0188 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.13.1.post200
- Datasets 2.9.0
- Tokenizers 0.13.2