feat: Adding support for Annbatch#3620
Open
ori-kron-wis wants to merge 56 commits intomainfrom
Open
Conversation
ori-kron-wis
added
on-merge: backport to 1.4.x
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #3620 +/- ##
==========================================
- Coverage 88.23% 88.08% -0.15%
==========================================
Files 229 229
Lines 22646 23373 +727
==========================================
+ Hits 19981 20589 +608
- Misses 2665 2784 +119
Flags with carried forward coverage won't be shown. Click here to find out more.
🚀 New features to boost your workflow:
|
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 6ee2098c75
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Collaborator
Author
|
Resolvi annabtch implementation still loads data into memory and can't be based on annbatch, although what was committed here. There is no real gain from it right now; need to think about graph dataloader for it as @canergen suggested. I will remove those commits from here and put them on a side branch for later use: Ori-annbatch-resolvi |
Added sample_key parameter to BaseModelClass.setup_annbatch to enable sample-level metadata tracking in annbatch datamodules. This is required for models like MrVI that operate on sample-level information. Also fixed import issues with anndata by updating to use experimental module for CSCDataset, CSRDataset, and read_elem. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
- Modified CellAssign.__init__ to accept registry= parameter - Added setup_annbatch method to compute streaming statistics (col_means, basis_means) - Modified train() to accept datamodule parameter - Added test_annbatch_setup_cellassign Note: Test fails due to missing size_factor field in annbatch datamodule This requires custom datamodule similar to ContrastiveVI/VELOVI 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
… into memory ad.experimental.read_lazy fails on h5ad files with zero-shape datasets in /uns (e.g. Decipher tutorial data with /uns/decipher/config/layers_z_to_x). Replace with direct h5py read of only the var group — reads index column via _index attr, with fallback to common names. Keeps data off memory per annbatch design. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
ilan-gold
reviewed
Apr 30, 2026
ori-kron-wis
added
custom_dataloader
…ation is done with same collection split
- _save_load.py: always override serialized datamodule string with actual datamodule arg so SCANVI and other models with datamodule param don't receive a string object at load time - _base_model.py: skip module recreation in setup_datamodule load path when no datamodule is provided (_initialize_model already builds correct module from registry); preserve library_log_means/vars buffers when recreating with datamodule - _training_mixin.py: handle adata=None + datamodule=None load path in _set_indices_and_labels by using categorical_mapping (without unlabeled category) so n_labels matches training-time computation - revert docs/tutorials/notebooks submodule to origin/main pointer Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…lidation performance with annbatch training
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
2 participants
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.
No description provided.