translaTHOR

translaTHOR is a language translation project focused on translating English text to Bengali. Leveraging advanced natural language processing techniques, it utilizes the Transformers library and PyTorch framework to build and train translation models. The project is currently under development.

Features

• English to Bengali Translation: Translate English sentences into Bengali using trained models.
• Model Training and Evaluation: Train models on the SUPARA-BENCHMARK dataset and evaluate their performance.
• Interactive Notebooks: Jupyter Notebooks provided for experimentation and demonstration purposes.

Technologies Used

• 🤗 Transformers Library
• PyTorch

Dataset

The project utilizes the SUPARA-BENCHMARK dataset for training and evaluation.

Resources

• Hugging Face Transformers Documentation
• BERT
• BERT
• Bangla BERT Base Model by Sagor Sarker
• Helsinki NLP Models
• OPUS

Getting Started

Prerequisites

Ensure you have the following installed:

• Python 3.x
• PyTorch
• Transformers
• Jupyter Notebook

Installation

1. Clone the repository:

   git clone https://github.com/PALLADIUM26/translaTHOR.git
   cd translaTHOR
   

2. Install dependencies:

   pip install torch transformers jupyter
   

3. Launch Jupyter Notebook:

   jupyter notebook
   

4. Open and run notebooks:

   • English_to_Bengali_0.ipynb
   • English_to_Bengali_1.ipynb
   • Translate_EnglishToBangla.ipynb

   These notebooks demonstrate the translation process and model usage.

Project Structure

• saved_model/: Directory containing saved model checkpoints.
• English_to_Bengali_0.ipynb: Jupyter Notebook for initial translation experiments.
• English_to_Bengali_1.ipynb: Continuation of translation experiments.
• Translate_EnglishToBangla.ipynb: Notebook demonstrating translation using trained models.
• english_to_bengali_0.py: Python script for translation tasks.
• translate_englishtobangla.py: Another Python script for translation tasks.
• README.md: Project documentation.

Contributing

Contributions are welcome! Feel free to fork the repository, make improvements, and submit pull requests. For major changes, please open an issue first to discuss your ideas.