gpt2_small_RU_superbpe_32768_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 4.7791
- Accuracy: 0.2767
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: 5e-05
- train_batch_size: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- lr_scheduler_warmup_steps: 1000
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|
| 7.6037 | 1.0 | 1252 | 6.4062 | 0.1832 |
| 6.2503 | 2.0 | 2504 | 5.6061 | 0.2205 |
| 5.671 | 3.0 | 3756 | 5.2716 | 0.2401 |
| 5.1861 | 4.0 | 5008 | 5.0933 | 0.2517 |
| 5.0476 | 5.0 | 6260 | 4.9801 | 0.2591 |
| 4.9498 | 6.0 | 7512 | 4.9025 | 0.2656 |
| 4.8761 | 7.0 | 8764 | 4.8470 | 0.2703 |
| 4.7708 | 8.0 | 10016 | 4.8113 | 0.2737 |
| 4.7248 | 9.0 | 11268 | 4.7887 | 0.2758 |
| 4.7006 | 10.0 | 12520 | 4.7791 | 0.2767 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
