<!-- 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. -->
Fake-News-Detection-Roberta
This model is a fine-tuned version of smallbenchnlp/roberta-small on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.0119
- Accuracy: 0.9981
- Precision: 0.9994
- Recall: 0.9966
- F1 score: 0.9980
- Auc score: 0.9980
Model description
The model is a fake news classifier which has been fined-tuned using the small version of pre-trained RoBERTa where Label 0 indicates its a Fake News but Label 1 indicates its a Real News.
Intended uses & limitations
To use the Inference API, please give input of word limit=70.
Training and evaluation data
The model was trained and tested using this dataset :https://www.kaggle.com/datasets/sadikaljarif/fake-news-detection-dataset-english
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 0.5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 score | Auc score |
|---|---|---|---|---|---|---|---|---|
| 0.0206 | 0.49 | 500 | 0.0119 | 0.9981 | 0.9994 | 0.9966 | 0.9980 | 0.9980 |
Framework versions
- Transformers 4.27.1
- Pytorch 1.13.1+cu116
- Tokenizers 0.13.2
