redit_gpt
This model use the architecture of gpt2 an was trained on the Ganz00/reddit-comments-cleaned-for-tg dataset. It achieves the following results on the evaluation set:
- Loss: 3.4867
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: 0.0005
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 5.2446 | 0.5891 | 1000 | 4.2133 |
| 3.9883 | 1.1782 | 2000 | 3.7935 |
| 3.7198 | 1.7673 | 3000 | 3.6230 |
| 3.5341 | 2.3564 | 4000 | 3.5221 |
| 3.4378 | 2.9455 | 5000 | 3.4867 |
Framework versions
- Transformers 4.42.4
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1