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

Fixes for large scale hyperopt #2083

Merged
merged 7 commits into from
Jun 7, 2022
Merged

Fixes for large scale hyperopt #2083

merged 7 commits into from
Jun 7, 2022

Conversation

ShreyaR
Copy link
Contributor

@ShreyaR ShreyaR commented Jun 2, 2022

  • Enable windowing for large datasets
  • Add support for tune.report while using ray backend
  • Add support for distributed hyperopt

if not enable_windowing:
pipe = self.ds.repeat()
else:
if window_size_bytes is None:
Copy link
Contributor Author

Choose a reason for hiding this comment

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

An alternative to requiring window_size_bytes to always be set is to calculate window size based on the following variables:

  • Total object store memory in the ray cluster, and
  • If running from within a tune trial and piping through the max concurrent trials parameter.
    Not sure if that would be overkill tho so I've implemented a simpler solution for now.

@ShreyaR ShreyaR requested a review from tgaddair June 2, 2022 08:26
@github-actions
Copy link

github-actions bot commented Jun 2, 2022

Unit Test Results

       6 files  ±0         6 suites  ±0   2h 25m 1s ⏱️ - 11m 15s
2 805 tests ±0  2 773 ✔️ ±0    32 💤 ±0  0 ±0 
8 415 runs  ±0  8 315 ✔️ ±0  100 💤 ±0  0 ±0 

Results for commit 11ee50b. ± Comparison against base commit d3eea13.

♻️ This comment has been updated with latest results.

ludwig/data/dataset/ray.py Outdated Show resolved Hide resolved
@ShreyaR ShreyaR closed this Jun 2, 2022
@ShreyaR ShreyaR reopened this Jun 2, 2022
@ShreyaR ShreyaR closed this Jun 2, 2022
@ShreyaR ShreyaR reopened this Jun 2, 2022
@ShreyaR ShreyaR closed this Jun 3, 2022
@ShreyaR ShreyaR reopened this Jun 3, 2022
@ShreyaR ShreyaR merged commit c35df1c into master Jun 7, 2022
@ShreyaR ShreyaR deleted the large-scale-hyperopt-v2 branch June 7, 2022 00:01
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.

None yet

2 participants