bertweet-emotion-base
This model is a fine-tuned version of Bertweet. It achieves the following results on the evaluation set:
- Loss: 0.1172
- Accuracy: 0.945
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 80
- eval_batch_size: 80
- lr_scheduler_type: linear
- num_epochs: 6.0
Framework versions
- Transformers 4.12.5
- Pytorch 1.10.0+cu113
- Datasets 1.15.1
- Tokenizers 0.10.3
- Downloads last month
- 25
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.
Dataset used to train Emanuel/bertweet-emotion-base
Spaces using Emanuel/bertweet-emotion-base 2
Evaluation results
- Accuracy on emotionself-reported0.945
- Accuracy on emotiontest set verified0.928
- Precision Macro on emotiontest set verified0.888
- Precision Micro on emotiontest set verified0.928
- Precision Weighted on emotiontest set verified0.929
- Recall Macro on emotiontest set verified0.886
- Recall Micro on emotiontest set verified0.928
- Recall Weighted on emotiontest set verified0.928
- F1 Macro on emotiontest set verified0.886
- F1 Micro on emotiontest set verified0.928