gpt2_small_FR_unigram_65536_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.3586
- Accuracy: 0.4002
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 |
|---|
| 5.8102 | 1.0 | 1567 | 4.0545 | 0.3473 |
| 3.86 | 2.0 | 3134 | 3.7394 | 0.3697 |
| 3.7142 | 3.0 | 4701 | 3.6004 | 0.3783 |
| 3.5529 | 4.0 | 6268 | 3.5189 | 0.3872 |
| 3.4938 | 5.0 | 7835 | 3.4619 | 0.3907 |
| 3.406 | 6.0 | 9402 | 3.4214 | 0.3949 |
| 3.3725 | 7.0 | 10969 | 3.3957 | 0.3973 |
| 3.3133 | 8.0 | 12536 | 3.3749 | 0.3988 |
| 3.2801 | 9.0 | 14103 | 3.3629 | 0.4003 |
| 3.2601 | 10.0 | 15670 | 3.3586 | 0.4002 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
