Google I/O

Large language models with Keras

The latest Keras 3 machine learning framework lets you write and run your code in JAX, Pytorch, or Tensorflow. Learn about Gemma, the large language model family of open models from Google. We will teach you basic and advanced LLM workflows, including chat generation, LoRA fine-tuning, model parallelism to train on large-scale infrastructure, style alignment, model surgery, and more.

Intermediate
Technical session
Join a community group

Meet developers, discover local groups, and build your global network.

Get started
Continue learning

Grow your skills around the Google technology you love.

Get started