| license: apache-2.0 | |
| tags: | |
| - generated_from_trainer | |
| model-index: | |
| - name: 20split_dataset | |
| 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. --> | |
| # 20split_dataset | |
| This model is a fine-tuned version of [bert-base-uncased](https://huggingface.co/bert-base-uncased) on the None dataset. | |
| It achieves the following results on the evaluation set: | |
| - Loss: 2.0446 | |
| ## 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: 64 | |
| - eval_batch_size: 8 | |
| - seed: 42 | |
| - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 | |
| - lr_scheduler_type: linear | |
| - num_epochs: 6 | |
| ### Training results | |
| | Training Loss | Epoch | Step | Validation Loss | | |
| |:-------------:|:-----:|:-----:|:---------------:| | |
| | 2.5971 | 1.0 | 11851 | 2.3479 | | |
| | 2.3773 | 2.0 | 23702 | 2.2446 | | |
| | 2.2663 | 3.0 | 35553 | 2.1630 | | |
| | 2.1842 | 4.0 | 47404 | 2.1059 | | |
| | 2.1145 | 5.0 | 59255 | 2.0626 | | |
| | 2.0652 | 6.0 | 71106 | 2.0446 | | |
| ### Framework versions | |
| - Transformers 4.20.1 | |
| - Pytorch 1.12.0+cu113 | |
| - Datasets 2.3.2 | |
| - Tokenizers 0.12.1 | |