Skip to content

fix runtime errors in colocated python dataloader #1819

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

Open
wants to merge 2 commits into
base: main
Choose a base branch
from

Conversation

sadikneipp
Copy link
Collaborator

@sadikneipp sadikneipp commented Jun 11, 2025

Title

Fix: Resolve runtime errors in colocated TFDS dataloader

Description

fixes:

  • _tfds_data_processing.py:

    • The tfds.builder now correctly uses the dataset_path variable, which is needed for tfds.
    • The constructor for RemoteIterator has been updated to pass global_shape as argument, fixing an interface mismatch.
  • multihost_dataloading.py:

    • The put_to_tpu_devices function now correctly returns the array after it has been placed on the TPU device. Otherwise the iterator would have empty data.
    • Initialization logging for RemoteIterator is now more informative, including a test output to aid in debugging and validation.

Tests

This patch was used in the last round of scale tests, validated up to 4 slice scale.

Checklist

Before submitting this PR, please make sure (put X in square brackets):

  • I have performed a self-review of my code.
  • I have necessary comments in my code, particularly in hard-to-understand areas.
  • I have run end-to-end tests and provided workload links above if applicable.
  • I have made or will make corresponding changes to the doc if needed.

Copy link
Collaborator

@aireenmei aireenmei left a comment

Choose a reason for hiding this comment

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

Thanks Sadi!

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