gpt2_small_FR_bpe_32768_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.9316
- Accuracy: 0.4427
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 |
|---|
| 5.1556 | 1.0 | 1788 | 3.5625 | 0.3859 |
| 3.4617 | 2.0 | 3576 | 3.3003 | 0.4055 |
| 3.2496 | 3.0 | 5364 | 3.1701 | 0.4183 |
| 3.1257 | 4.0 | 7152 | 3.0863 | 0.4267 |
| 3.0692 | 5.0 | 8940 | 3.0316 | 0.4315 |
| 2.9937 | 6.0 | 10728 | 2.9942 | 0.4357 |
| 2.9414 | 7.0 | 12516 | 2.9663 | 0.4394 |
| 2.9037 | 8.0 | 14304 | 2.9478 | 0.4412 |
| 2.8738 | 9.0 | 16092 | 2.9371 | 0.4427 |
| 2.855 | 10.0 | 17880 | 2.9316 | 0.4427 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.9.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
