Ziad Helal ZiadHelal

Hi there, I'm Ziad Helal 😃

I'm currently a Master's student in Computer Science at University of Freiburg - Faculty of Engineering.
I have 1+ years of experience building and deploying machine learning and deep learning models.
Exploring the world of Large Language Models (LLMs) and Generative AI—always learning, always building. I'm diving deep into the algorithms, architectures, and cutting-edge techniques that power these technologies, with a goal to turn theory into practice and create innovative AI solutions. Let’s build the future of AI, one model at a time.

MS (ongoing) | Computer Science (AI Specialization) | University of Freiburg
BS (2021) | Computer Engineering | Cairo University

Quantization for VeRA using bitsandbytes 🚀
Added 4-bit and 8-bit quantization to the VeRA method, leveraging bitsandbytes.

Key Features:
- 🛠️ New bnb.py for 4-bit & 8-bit layers.
- ⚡ Enhanced VeRA for lightning-fast, memory-efficient inference.
- ✅ Added unit tests for robust quantization support.
- 📚 Updated documentation to reflect cutting-edge capabilities.
Impact: Empowered Huggingface's PEFT library with state-of-the-art quantized inference for efficient and scalable AI models.