| tags: | |
| - generated_from_trainer | |
| model-index: | |
| - name: nezha-chinese-base-finetuned-product | |
| 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. --> | |
| # nezha-chinese-base-finetuned-product | |
| This model is a fine-tuned version of [peterchou/nezha-chinese-base](https://huggingface.co/peterchou/nezha-chinese-base) on the None dataset. | |
| It achieves the following results on the evaluation set: | |
| - Loss: 0.0004 | |
| ## 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: 2e-05 | |
| - 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: linear | |
| - num_epochs: 3.0 | |
| ### Training results | |
| | Training Loss | Epoch | Step | Validation Loss | | |
| |:-------------:|:-----:|:-----:|:---------------:| | |
| | 0.0309 | 1.0 | 6473 | 0.0037 | | |
| | 0.0033 | 2.0 | 12946 | 0.0006 | | |
| | 0.0017 | 3.0 | 19419 | 0.0004 | | |
| ### Framework versions | |
| - Transformers 4.17.0 | |
| - Pytorch 1.6.0 | |
| - Datasets 2.0.0 | |
| - Tokenizers 0.11.6 | |