metadata
base_model: cardiffnlp/twitter-roberta-base-sentiment-latest
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: undersampled-review-clf
results: []
datasets:
- justina/yelp_boba_reviews
undersampled-review-clf
This model is a fine-tuned version of cardiffnlp/twitter-roberta-base-sentiment-latest on justina/yelp-boba-reviews dataset. Undersampling techniques were used to optimize the model for predicting Yelp review ratings.
It achieves the following results on the evaluation set:
- Loss: 0.4412
- F1 Macro: 0.7799
- Aucpr Macro: 0.8286
- Accuracy: 0.8464
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: 1e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | F1 Macro | Aucpr Macro | Accuracy |
|---|---|---|---|---|---|---|
| 0.9348 | 1.22 | 100 | 0.7286 | 0.6132 | 0.6244 | 0.6962 |
| 0.7438 | 2.44 | 200 | 0.7857 | 0.6232 | 0.6215 | 0.6735 |
| 0.6275 | 3.66 | 300 | 0.8317 | 0.5976 | 0.6092 | 0.6778 |
| 0.5561 | 4.88 | 400 | 0.8176 | 0.6200 | 0.6238 | 0.6868 |
Framework versions
- Transformers 4.32.1
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3