<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
gpt2-cl-length-sampling
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 5.0276
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
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 1
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
6.4896 | 0.1 | 500 | 5.9335 |
5.1803 | 0.2 | 1000 | 5.5576 |
4.871 | 0.3 | 1500 | 5.3672 |
4.6547 | 0.4 | 2000 | 5.2453 |
4.5086 | 0.5 | 2500 | 5.1611 |
4.3642 | 0.6 | 3000 | 5.0948 |
4.2412 | 0.7 | 3500 | 5.0350 |
4.1326 | 0.8 | 4000 | 4.9978 |
4.0612 | 0.9 | 4500 | 4.9717 |
4.0255 | 1.0 | 5000 | 4.9665 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3