AI-Hypercomputer / AI-Hypercomputer/maxtext

Make maxtext_xpk_runner support other trainers

Ouverte
#1,552 0 commentaires 0 réactions 2 personnes assignées Réclamée par @suexu1025 Voir sur GitHub
feature request
Langage dominant
Python
Étoiles
2.4k
Forks
607
Merge moyen
2 j 19 h
PR mergées (30 j)
158

Description

At the moment, the command is [hard coded](https://github.com/AI-Hypercomputer/maxtext/blob/1e4d513ad70dd4074d975a9f7936295008d4b900/benchmarks/maxtext_xpk_runner.py#L412C20-L412C33) to execute the `MaxText.train` module. If the module to execute could be configured in the WorkloadConfig, that would allow for additional trainers being developed to make use of the benchmark runner.

I am running into this while wanting to test a new trainer `MaxText.elastic_train` in PR #1310 but there are also other trainers being developed and merged and I think configuring `maxtext_xpk_runner` will be useful moving forward.

Guide de contribution

Ouvrir le guide de contribution

Évaluation

Cette issue n'a pas encore été évaluée.

Recevez les nouvelles issues par e-mail

Un résumé court des issues GitHub adaptées aux débutants.