Fine-tuning LLMs with Declarative ML Orchestration

This talk is part of the AI Engineering Job Fair. Signing up for this event will give you access to all the talks on the agenda, including the job fair.

About the talk

Foundation LLMs are trained by a few organizations with massive compute resources. The ML community fine-tunes these models for specific uses, facing infrastructure challenges.

In this session, Niels will demonstrate how to use Flyte, a Linux Foundation open-source orchestration platform.

Flyte allows for the declarative specification of the infrastructure needed for a broad range of ML workloads, including fine-tuning LLMs with limited resources.

Key Takeaways

  • Efficiently fine-tune large language models with limited resources using Flyte.
  • Explore torchrun, LoRA, and FSDP for deep learning tasks.
  • Leverage Flyte's reproducibility and cost management features for ML workloads.
Job Fair

About the speaker

Niels is the Chief Machine Learning Engineer at Union.ai, a core maintainer of Flyte, and creator of UnionML and Pandera. His mission is to help data science and machine learning practitioners be more productive while doing research in the ML space.

High-income, remote careers for developers like you

Join Arc and receive offers from high-growth tech startups paying from $60,000 to 175,000 USD!

Discussion 

Loading...
By using Codementor, you agree to our Cookie Policy.