trocr-barbados-fold3
This model is a fine-tuned version of microsoft/trocr-base-handwritten on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.9763
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: 3e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 15
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 1.0 | 205 | 1.3033 |
| No log | 2.0 | 410 | 1.1299 |
| 3.0802 | 3.0 | 615 | 1.0688 |
| 3.0802 | 4.0 | 820 | 1.0312 |
| 1.1270 | 5.0 | 1025 | 1.0424 |
| 1.1270 | 6.0 | 1230 | 1.0219 |
| 1.1270 | 7.0 | 1435 | 1.0208 |
| 0.5213 | 8.0 | 1640 | 0.9953 |
| 0.5213 | 9.0 | 1845 | 0.9983 |
| 0.2403 | 10.0 | 2050 | 0.9864 |
| 0.2403 | 11.0 | 2255 | 0.9980 |
| 0.2403 | 12.0 | 2460 | 0.9879 |
| 0.1119 | 13.0 | 2665 | 0.9917 |
| 0.1119 | 14.0 | 2870 | 0.9841 |
| 0.0567 | 15.0 | 3075 | 0.9763 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2