gpt2_medium_AR_bpe_65536_parallel3-100_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 3.5902
- Accuracy: 0.3925
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: 64
- eval_batch_size: 64
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- total_train_batch_size: 128
- total_eval_batch_size: 128
- 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 |
|---|
| 4.266 | 1.0 | 7193 | 4.1474 | 0.3390 |
| 3.9434 | 2.0 | 14386 | 3.8723 | 0.3606 |
| 3.7727 | 3.0 | 21579 | 3.7565 | 0.3716 |
| 3.6591 | 4.0 | 28772 | 3.6892 | 0.3781 |
| 3.5685 | 5.0 | 35965 | 3.6465 | 0.3835 |
| 3.4862 | 6.0 | 43158 | 3.6200 | 0.3864 |
| 3.4171 | 7.0 | 50351 | 3.6033 | 0.3889 |
| 3.3485 | 8.0 | 57544 | 3.5963 | 0.3906 |
| 3.2921 | 9.0 | 64737 | 3.5900 | 0.3919 |
| 3.2461 | 10.0 | 71930 | 3.5902 | 0.3925 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
