gpt2 persian question answering

Introduction

The GPT2-Persian-Question-Answering model is developed as a project within the Flax/JAX Community Week, organized by Hugging Face with TPU resources sponsored by Google. It is designed for question-answering tasks in Persian, utilizing the GPT-2 architecture.

Architecture

This model is based on the GPT-2 architecture, which is popular for text generation tasks. The implementation is compatible with libraries like PyTorch, TensorFlow, and JAX, and supports text-generation inference.

Training

The model is trained on the PersianQA dataset, a reading comprehension dataset derived from Persian Wikipedia. This dataset provides the necessary Persian language context to fine-tune the GPT-2 model for question-answering tasks.

Guide: Running Locally

1. Environment Setup: Ensure you have Python installed along with package managers like pip or conda.
2. Library Installation: Use pip to install Hugging Face Transformers and other dependencies:

   pip install transformers torch
   

3. Model Download: Download the model from Hugging Face's model hub.
4. Inference: Use the model with a simple script to input questions and receive answers in Persian.

For optimal performance, especially during training, consider using cloud-based GPU services such as AWS, Google Cloud, or Azure.

License

The model and its associated datasets and code are likely under licenses that permit research and development use. Please refer to the model card on Hugging Face for specific licensing details.