Learn Your Codebase: Fine-tuning CodeLlama with Flyte to Learn Flyte

Discover how to fine-tune a Large Language Model (LLM) on a specific codebase using Flyte, an open-source orchestration platform. In this conference talk, Niels Bantilan, Chief ML Engineer at Union.ai, demonstrates the process of fine-tuning CodeLlama on the Flyte codebase itself. Explore the challenges of provisioning appropriate infrastructure for fine-tuning LLMs and learn how Flyte's declarative specification capabilities can overcome these obstacles. Gain insights into leveraging multi-node, multi-gpu distributed training for efficient fine-tuning with limited resources. Understand the broader context of foundation LLMs, their training requirements, and the role of fine-tuning in adapting these models for specific applications.