gpt2_small_AR_unigram_49152_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 4.4877
- Accuracy: 0.3315
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 |
|---|
| 6.7928 | 1.0 | 1426 | 5.4846 | 0.2790 |
| 5.5222 | 2.0 | 2852 | 5.0159 | 0.3020 |
| 4.9559 | 3.0 | 4278 | 4.8214 | 0.3118 |
| 4.8239 | 4.0 | 5704 | 4.7026 | 0.3174 |
| 4.6613 | 5.0 | 7130 | 4.6272 | 0.3221 |
| 4.5917 | 6.0 | 8556 | 4.5724 | 0.3251 |
| 4.5455 | 7.0 | 9982 | 4.5367 | 0.3274 |
| 4.4627 | 8.0 | 11408 | 4.5107 | 0.3298 |
| 4.4385 | 9.0 | 12834 | 4.4941 | 0.3312 |
| 4.3969 | 10.0 | 14260 | 4.4877 | 0.3315 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
