Description:

InstructLab is a model-agnostic open-source initiative that invites everyone to partake in shaping the future of generative AI. Our goal is simple: make it possible for anyone to enhance Large Language Models (LLMs) with their contributions, breaking down the barriers that have traditionally restricted such advancements to those with extensive resources. This BoF session will delve into the innovative LAB method, a collaborative effort born from the MIT-IBM Watson AI Lab, which has dramatically simplified the model training process, reducing the reliance on large-scale human-generated data and intensive computing power.

Join us in the BoF to discover how you can become an active participant in the InstructLab community, contributing to the ongoing evolution of AI. This session is designed to demystify the complexities of traditional LLM training methods by contrasting them with InstructLab’s more accessible approach. You’ll learn how just a few human-generated examples can be transformed into a wealth of synthetic data, propelling the tuning and enhancement of models in ways previously thought impractical. We’ll also explore the synergy between InstructLab and retrieval-augmented generation (RAG), showcasing how these tools together can supercharge AI capabilities. See how InstructLab is not just about building smarter AI, but about fostering a more inclusive and collaborative AI development environment. we’re actively contributing to its growth and making it work for everyone. Join this journey and see how your contributions can make a real impact

Prerequisites:

pip install instructlab

Video URL:

https://youtu.be/CnHBjDfiWjo?feature=shared

Content URLs:

Project repo on Github : https://github.com/instructlab/instructlab HuggingFace : https://huggingface.co/instructlab Open Community DataSet Taxonomy : https://github.com/instructlab/taxonomy InstructLab : https://instructlab.ai/

Speaker Info:

Ramakrishna Reddy Yekulla ("Ramky") serves as a Principal Architect within the Application Developer Business Unit at Red Hat. In this role, he focuses on creating distinctive platform features that enhance developer efficiency, streamline data and application connections, and empower customers with the knowledge and resources to prioritize software supply chain security, minimize technical obligations, and achieve compliance goals. His interests lies in System Design, Functional Programming and Observability.

Speaker Links:

Open Hybrid Cloud & AI Developer Experience Keynote Talk at GIDS - https://youtu.be/KNGvtEVd5xI?feature=shared Simplifying Kubernetes: Streamlining Secure App Deployment in DevSecOps - https://youtu.be/bsAqoWiGmpQ?feature=shared Safer Driver Human Driver Perception Platform - https://youtu.be/qNcj2I9vhmw?feature=shared Driving developer productivity - https://youtu.be/sbHKyP-mEhU?feature=shared