trackmania-tmnf-distilgpt2-v2
This model is a fine-tuned version of distilgpt2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.2489
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 8.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.6429 | 0.1095 | 200 | 1.5249 |
| 1.5576 | 0.2189 | 400 | 1.4750 |
| 1.5404 | 0.3284 | 600 | 1.4393 |
| 1.5029 | 0.4379 | 800 | 1.4116 |
| 1.4252 | 0.5473 | 1000 | 1.3929 |
| 1.4361 | 0.6568 | 1200 | 1.3649 |
| 1.3958 | 0.7663 | 1400 | 1.3523 |
| 1.4064 | 0.8758 | 1600 | 1.3419 |
| 1.3853 | 0.9852 | 1800 | 1.3217 |
| 1.3707 | 1.0947 | 2000 | 1.3137 |
| 1.3580 | 1.2042 | 2200 | 1.3013 |
| 1.3785 | 1.3136 | 2400 | 1.2921 |
| 1.3552 | 1.4231 | 2600 | 1.2855 |
| 1.3694 | 1.5326 | 2800 | 1.2751 |
| 1.3128 | 1.6420 | 3000 | 1.2692 |
| 1.3424 | 1.7515 | 3200 | 1.2673 |
| 1.3293 | 1.8610 | 3400 | 1.2614 |
| 1.3435 | 1.9704 | 3600 | 1.2524 |
| 1.2980 | 2.0799 | 3800 | 1.2489 |
Framework versions
- Transformers 5.14.1
- Pytorch 2.13.0+cu130
- Datasets 5.0.1
- Tokenizers 0.22.2