ocr-captcha-v3-finetuned
This model is a fine-tuned version of anuashok/ocr-captcha-v3 on the dathuynh1108/captcha-to-text dataset. It achieves the following results on the evaluation set:
- Loss: 0.1145
- Cer: 0.0056
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: 1e-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: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 0.0563 | 1.0 | 2532 | 0.0877 | 0.0074 |
| 0.0144 | 2.0 | 5064 | 0.1097 | 0.008 |
| 0.0175 | 3.0 | 7596 | 0.1212 | 0.0068 |
| 0.0210 | 4.0 | 10128 | 0.1025 | 0.0071 |
| 0.0007 | 5.0 | 12660 | 0.1109 | 0.0068 |
| 0.0003 | 6.0 | 15192 | 0.1112 | 0.0057 |
| 0.0023 | 7.0 | 17724 | 0.1164 | 0.0060 |
| 0.0001 | 8.0 | 20256 | 0.1117 | 0.0056 |
| 0.0000 | 9.0 | 22788 | 0.1153 | 0.0054 |
| 0.0000 | 10.0 | 25320 | 0.1145 | 0.0056 |
Framework versions
- Transformers 5.8.0
- Pytorch 2.11.0+cu130
- Datasets 4.8.5
- Tokenizers 0.22.2