Dataset Specific Splitting#18
Open
MattsonCam wants to merge 2 commits into
Open
Conversation
added 2 commits
May 14, 2026 10:39
nuclear speckle data
Add a per-dataset holdout plate configuration in train.py and filter manifest rows before building CellCropToCropDataset so held-out plates never reach HashSplitter. This keeps holdout data in cache while preventing it from being used in train/val splits. Also log holdout plate and pre/post filter row counts to MLflow, and fail fast if filtering removes all manifest rows.
This file contains hidden or 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This pr updates train.py to exclude a dataset-specific holdout plate before building train/val/test splits, so evaluation is done on data not seen during split generation. It also logs holdout filtering counts in MLflow and slightly adjusts split proportions to keep the test share aligned after filtering. Holding out a plate reduces the available data by at least 12% , which is why I decided to not also holdout wells from the same plate at the moment.