gpt2_small_EN_unigram_65536_parallel3-100_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1488
- Accuracy: 0.4246
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 |
|---|
| 3.8033 | 1.0 | 7523 | 3.6787 | 0.3746 |
| 3.5061 | 2.0 | 15046 | 3.4419 | 0.3941 |
| 3.3843 | 3.0 | 22569 | 3.3385 | 0.4040 |
| 3.3016 | 4.0 | 30092 | 3.2775 | 0.4101 |
| 3.2424 | 5.0 | 37615 | 3.2364 | 0.4140 |
| 3.1963 | 6.0 | 45138 | 3.2057 | 0.4175 |
| 3.1548 | 7.0 | 52661 | 3.1838 | 0.4201 |
| 3.1236 | 8.0 | 60184 | 3.1679 | 0.4219 |
| 3.0973 | 9.0 | 67707 | 3.1546 | 0.4238 |
| 3.074 | 10.0 | 75230 | 3.1488 | 0.4246 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
