gpt2_small_FI_unigram_32768_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.8602
- Accuracy: 0.3978
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 |
|---|
| 6.1151 | 1.0 | 1422 | 4.8389 | 0.3336 |
| 4.8462 | 2.0 | 2844 | 4.3943 | 0.3587 |
| 4.3122 | 3.0 | 4266 | 4.1987 | 0.3716 |
| 4.1828 | 4.0 | 5688 | 4.0843 | 0.3788 |
| 4.0247 | 5.0 | 7110 | 4.0056 | 0.3849 |
| 3.9573 | 6.0 | 8532 | 3.9507 | 0.3891 |
| 3.9108 | 7.0 | 9954 | 3.9120 | 0.3920 |
| 3.8331 | 8.0 | 11376 | 3.8850 | 0.3951 |
| 3.8053 | 9.0 | 12798 | 3.8670 | 0.3971 |
| 3.7664 | 10.0 | 14220 | 3.8602 | 0.3978 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
