GoogleCloudPlatform / GoogleCloudPlatform/kubernetes-engine-samples
Adding tutorial content for "Train a Large Language Model (LLM) using TPUs on GKE with JAX"
Open
- Dominant language
- HCL
- Stars
- 1.4k
- Forks
- 1.3k
- Avg merge
- 2d 2h
- Merged PRs (30d)
- 4
Description
Propose to adding content (Scripts and training code files) for the tutorial [Train a Large Language Model (LLM) using TPUs on GKE with JAX](https://docs.google.com/document/d/1Tl79ct-KWPgdytWCSIsR_Q_ZDDdjZCTjUbQVSckhDrs/edit?resourcekey=0-oIOsGt08A0ZCXwkx7DTDZw&tab=t.0#heading=h.39htcwzd6s73),
The changes have been validated and tested on GKE for both standard and autopilot clusters.
CL for the current draft tutorial: https://critique.corp.google.com/cl/855858604
Contributor guide
Assessment
This issue has not been assessed yet.