gpt2_small_FR_superbpe_49152_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 4.6532
- Accuracy: 0.2633
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 |
|---|
| 7.5036 | 1.0 | 1160 | 6.0292 | 0.1932 |
| 5.8056 | 2.0 | 2320 | 5.3407 | 0.2200 |
| 5.3421 | 3.0 | 3480 | 5.0884 | 0.2334 |
| 5.1186 | 4.0 | 4640 | 4.9356 | 0.2426 |
| 4.9698 | 5.0 | 5800 | 4.8348 | 0.2489 |
| 4.8608 | 6.0 | 6960 | 4.7661 | 0.2536 |
| 4.7043 | 7.0 | 8120 | 4.7160 | 0.2581 |
| 4.6499 | 8.0 | 9280 | 4.6820 | 0.2610 |
| 4.6099 | 9.0 | 10440 | 4.6608 | 0.2628 |
| 4.5802 | 10.0 | 11600 | 4.6532 | 0.2633 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
