Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Distributed ml on k8s - first tutorial #280

Merged
merged 21 commits into from
Jan 9, 2025
Merged

Distributed ml on k8s - first tutorial #280

merged 21 commits into from
Jan 9, 2025

Conversation

matbun
Copy link
Collaborator

@matbun matbun commented Jan 7, 2025

This is somehow related to #172, but introduces distributed ML on k8s using Kubeflow's training operator, rather than focusing on pipelines or ML tracking.

Other changes:

  • Fix some problems in the pyproject.toml, which now should be a bit more stable (updated uv lock accordingly).
  • Updated the TorchTrainer and TorchDistributedStrategy to support gloo backend for CPU-only distributed training.
  • Added support for tqdm progress bars in the trainer (can be disabled by the user)... Not a fan of tqdm, but it is useful when running k8s pods.

Future work include distributed ML on GPU nodes and k8s persistent volumes.

Copy link
Collaborator

@jarlsondre jarlsondre left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@matbun matbun merged commit 3fc10ef into main Jan 9, 2025
13 checks passed
@matbun matbun deleted the dist-ml-k8s branch January 9, 2025 15:27
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants