metadata
license: apache-2.0
tags:
- generated_from_trainer
metrics:
- accuracy
- f1
- precision
- recall
model-index:
- name: FakevsRealNews
results: []
Coding challenge
The challenge involved building a fake news classifier using the huggingface library.
This final model is a fine-tuned version of distilbert-base-uncased on an fake-and-real-news dataset. The link to the dataset is https://www.kaggle.com/datasets/clmentbisaillon/fake-and-real-news-dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0000
- Accuracy: 1.0
- F1: 1.0
- Precision: 1.0
- Recall: 1.0
Model description
Finetuned Distilbert
Training and evaluation data
The training data was split into train-dev-test in the ratio 80-10-10.
Training procedure
The title and text of each news story was concatenated to form each datapoint. Then a model was finetuned to perform single label classification on each datapoint. The final prediction is the class with the highest probability.
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
|---|---|---|---|---|---|---|---|
| 0.0503 | 1.0 | 1956 | 0.0025 | 0.9995 | 0.9995 | 0.9995 | 0.9995 |
| 0.001 | 2.0 | 3912 | 0.0001 | 1.0 | 1.0 | 1.0 | 1.0 |
| 0.0007 | 3.0 | 5868 | 0.0000 | 1.0 | 1.0 | 1.0 | 1.0 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.0+cu111
- Datasets 2.1.0
- Tokenizers 0.12.1