gpt2_small_FR_unigram_8192_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 2.8809
- Accuracy: 0.4311
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 |
|---|
| 5.0887 | 1.0 | 1844 | 3.5564 | 0.3490 |
| 3.4588 | 2.0 | 3688 | 3.2489 | 0.3818 |
| 3.2091 | 3.0 | 5532 | 3.1071 | 0.3989 |
| 3.0807 | 4.0 | 7376 | 3.0313 | 0.4093 |
| 3.0006 | 5.0 | 9220 | 2.9797 | 0.4165 |
| 2.9424 | 6.0 | 11064 | 2.9434 | 0.4204 |
| 2.9047 | 7.0 | 12908 | 2.9167 | 0.4241 |
| 2.8651 | 8.0 | 14752 | 2.8995 | 0.4275 |
| 2.836 | 9.0 | 16596 | 2.8874 | 0.4291 |
| 2.8124 | 10.0 | 18440 | 2.8809 | 0.4311 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
