| tags: | |
| - generated_from_trainer | |
| model-index: | |
| - name: GPT-Neo_DnD | |
| results: [] | |
| <!-- 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. --> | |
| # GPT-Neo_DnD | |
| This model is a fine-tuned version of [PatrickTyBrown/GPT-Neo_DnD](https://huggingface.co/PatrickTyBrown/GPT-Neo_DnD) on the None dataset. | |
| It achieves the following results on the evaluation set: | |
| - eval_loss: 3.3503 | |
| - eval_runtime: 88.7855 | |
| - eval_samples_per_second: 6.51 | |
| - eval_steps_per_second: 3.255 | |
| - epoch: 1.97 | |
| - step: 11400 | |
| ## 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: 2.5e-05 | |
| - train_batch_size: 2 | |
| - eval_batch_size: 2 | |
| - seed: 42 | |
| - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 | |
| - lr_scheduler_type: linear | |
| - num_epochs: 3 | |
| ### Framework versions | |
| - Transformers 4.21.3 | |
| - Pytorch 1.12.1+cu113 | |
| - Datasets 2.4.0 | |
| - Tokenizers 0.12.1 | |