Simple BERT Text Classification Model

This repository provides a simple implementation of a text classification model using BERT (Bidirectional Encoder Representations from Transformers).

Features

• Uses Hugging Face's Transformers library.
• Simple and easy-to-understand implementation.
• Includes scripts for data preprocessing, model training, and prediction.

Requirements

• Python 3.6+
• PyTorch
• Transformers
• scikit-learn
• pandas

Install the required packages using:

pip install -r requirements.txt

Usage

1. Clone the Repository

git clone https://github.com/StarJulian/Simple-BERT-text-classification-model.git
cd Simple-BERT-text-classification-model

2. Prepare Your Dataset

Prepare your dataset in CSV format with text and label columns.

3. Train the Model

python train.py 

4. Evaluate the Model

python test.py 

5. Make Predictions

python predict.py

Repository Structure

Simple-BERT-text-classification-model/
├── datasets/
│   ├── your_train_dataset.txt
│   ├── your_testt_dataset.txt
├── saved_models/
│   └── your_model.pth
├── datasets.py
├── main.py
├── predict.py
├── requirements.txt
├── test.py
├── train.py
└── README.md

Scripts

• datasets.py: Script for data preprocessing.
• train.py: Script for training the BERT model.
• test.py: Script for evaluating the model.
• predict.py: Script for making predictions on new data.

License

This project is licensed under the Apache-2.0 License. See the LICENSE file for details.

Acknowledgements

• Hugging Face Transformers
• PyTorch