For sample indexing we fix the uneven sampling #226
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Collaborator
hariharan-devarajan
commented
Aug 30, 2024
hariharan-devarajan
commented
- Fix uneven sampling done for index-based and iterative
- Add a validation step to ensure we can validate that global indices are correctly shuffled and no indices are lost.
- Make sure we do file and sample shuffling in the reconfigure step.
- Remove sample shuffling from the dataloader Sampler code.
- Added test case to support uneven file distributions Fixes IndexError: list index out of range when running custom.yaml file with custom num_files_train parameter #225
hariharan-devarajan
force-pushed
the
bugfix/sample_indexing
branch
from
9160722
to
a535b0d
Compare
1. Fix uneven sampling done for index based and iterative 2. Add a validation step to ensure we can validate that global indices are correctly shuffled and no indices are lost. 3. Make sure we do file and sample shuffling in reconfigure step. 4. Remove sample shuffling from dataloader Sampler code. 5. Added test case to support uneven file distributions #225
hariharan-devarajan
force-pushed
the
bugfix/sample_indexing
branch
from
a535b0d
to
adf0649
Compare
hariharan-devarajan
force-pushed
the
bugfix/sample_indexing
branch
from
4ee8f09
to
6b2f3da
Compare
|
@zhenghh04 I investigated #221. It was not redundant shuffling. The current shuffling on reconfigure was only doing file shuffling, and the sampler in data loaders was doing sample shuffling. I made it more streamlined by doing both during reconfigure. Also added a validation step on reconfigure to make sure we read everything in every epoch. This PR Fixes #221 |
|
@zhenghh04 Can we merge this? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.