gpt2_small_FI_unigram_5388_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.1420
- Accuracy: 0.4221
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 |
|---|
| 5.4442 | 1.0 | 1746 | 4.0077 | 0.3224 |
| 3.8301 | 2.0 | 3492 | 3.5812 | 0.3660 |
| 3.5177 | 3.0 | 5238 | 3.4134 | 0.3861 |
| 3.423 | 4.0 | 6984 | 3.3236 | 0.3981 |
| 3.3022 | 5.0 | 8730 | 3.2603 | 0.4060 |
| 3.2282 | 6.0 | 10476 | 3.2173 | 0.4112 |
| 3.1726 | 7.0 | 12222 | 3.1871 | 0.4153 |
| 3.1405 | 8.0 | 13968 | 3.1642 | 0.4184 |
| 3.0982 | 9.0 | 15714 | 3.1492 | 0.4200 |
| 3.0741 | 10.0 | 17460 | 3.1420 | 0.4221 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
