gpt2_small_PL_unigram_81920_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.9089
- Accuracy: 0.4029
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: 95
- 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 |
|---|
| 6.3438 | 1.0 | 1868 | 4.7713 | 0.3478 |
| 4.6437 | 2.0 | 3736 | 4.3874 | 0.3701 |
| 4.3318 | 3.0 | 5604 | 4.2173 | 0.3800 |
| 4.1618 | 4.0 | 7472 | 4.1132 | 0.3864 |
| 4.0455 | 5.0 | 9340 | 4.0389 | 0.3924 |
| 3.9619 | 6.0 | 11208 | 3.9882 | 0.3953 |
| 3.8961 | 7.0 | 13076 | 3.9551 | 0.3983 |
| 3.8497 | 8.0 | 14944 | 3.9294 | 0.4012 |
| 3.8063 | 9.0 | 16812 | 3.9144 | 0.4022 |
| 3.777 | 10.0 | 18680 | 3.9089 | 0.4029 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
