gpt2_small_KO_unigram_8192_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.7410
- Accuracy: 0.3511
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 |
|---|
| 5.9144 | 1.0 | 1641 | 4.5428 | 0.2845 |
| 4.3633 | 2.0 | 3282 | 4.1745 | 0.3101 |
| 4.1913 | 3.0 | 4923 | 4.0165 | 0.3237 |
| 3.9997 | 4.0 | 6564 | 3.9205 | 0.3321 |
| 3.8921 | 5.0 | 8205 | 3.8608 | 0.3383 |
| 3.8417 | 6.0 | 9846 | 3.8166 | 0.3419 |
| 3.7748 | 7.0 | 11487 | 3.7839 | 0.3460 |
| 3.7293 | 8.0 | 13128 | 3.7627 | 0.3477 |
| 3.6988 | 9.0 | 14769 | 3.7471 | 0.3503 |
| 3.6702 | 10.0 | 16410 | 3.7410 | 0.3511 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
