Let every layer take its initializers - #78
Merged
jessegrabowski merged 6 commits intoAug 10, 2026
Conversation
Passes as written: it pins the behavior gh pymc-devs#70 established so the initializer keywords landing next cannot quietly strip a declaration, which is the one way they break something.
Its two arguments both default, which is what lets it into a registry whose entries are built with no arguments; a differently parameterized initializer can only be passed as an instance.
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #78 +/- ##
==========================================
+ Coverage 95.43% 95.46% +0.02%
==========================================
Files 45 45
Lines 1798 1809 +11
==========================================
+ Hits 1716 1727 +11
Misses 82 82 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
jessegrabowski
added
enhancement
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 layer accepted an initializer, so changing how a parameter was drawn meant poking
param.initializerbetween construction andModel.initialize()— undocumented, and silently reverted if you did it the other way round. Every leaf layer now takes one per parameter it owns, keyword-only, setting the declaration mechanism from #70 rather than adding a second source of truth. Omitting a keyword changes nothing:Linear's bias still starts at zeros and the norms at the identity transform.A
NormalInitializercomes with it, which the registry was missing. Composed layers are untouched on purpose; threading a keyword throughAttentionandTransformerBlockfor per-layer scaled init is next.