gpt2_small_AR_unigram_32768_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 4.3262
- Accuracy: 0.3328
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 |
|---|
| 6.598 | 1.0 | 1480 | 5.2864 | 0.2771 |
| 5.357 | 2.0 | 2960 | 4.8393 | 0.3014 |
| 4.8015 | 3.0 | 4440 | 4.6499 | 0.3111 |
| 4.6774 | 4.0 | 5920 | 4.5361 | 0.3183 |
| 4.5105 | 5.0 | 7400 | 4.4614 | 0.3224 |
| 4.4527 | 6.0 | 8880 | 4.4112 | 0.3257 |
| 4.3602 | 7.0 | 10360 | 4.3737 | 0.3289 |
| 4.3262 | 8.0 | 11840 | 4.3481 | 0.3311 |
| 4.2722 | 9.0 | 13320 | 4.3334 | 0.3320 |
| 4.2524 | 10.0 | 14800 | 4.3262 | 0.3328 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
