model_output_original_subreddit-wallstreetbets_1
This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 3.5436
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: 64
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 512
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 2
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 3.8943 | 0.25 | 1000 | 3.8122 |
| 3.799 | 0.5 | 2000 | 3.7199 |
| 3.7425 | 0.75 | 3000 | 3.6688 |
| 3.6938 | 1.0 | 4000 | 3.6269 |
| 3.543 | 1.25 | 5000 | 3.5972 |
| 3.5417 | 1.5 | 6000 | 3.5657 |
| 3.5122 | 1.75 | 7000 | 3.5477 |
| 3.4857 | 1.99 | 8000 | 3.5436 |
Framework versions
- Transformers 4.24.0
- Pytorch 1.13.0+cu117
- Datasets 2.7.1
- Tokenizers 0.13.2