Intel OneAPI - LLM

This project is a showcase to a solution for "oneAPI Hackathon: The LLM Challenge" using BERT models from Hugging Face.

💾 Dataset

The dataset for this challenge can be accessed from here.

📚 Content

• Jupyter Notebooks
• Web app code

⚙️ Models

The models can be accessed from Hugging Face -

• Extractive QA model
• Yes/No QA model
• True/False QA model

💻 System info

The Jupyter notebooks are excuted on Intel® Developer Cloud (Intel Xeon CPU with single Intel Max 1100 GPU) using tensorflow_xpu environment.

🕹️ Demo

Note

To get True/False answer on the web app, the question needs to be tagged with the prefix "true or false:".

🔗 Links