fix: allow JAX 0.11 downstream environments - #802
Merged
Merged
Conversation
Refresh only JAX and JAXlib in the lock and benchmark constraints. Co-authored-by: OpenAI Codex <codex@openai.com>
Collaborator
|
Thanks @AHMETHAKANBEZIR1 ! |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Fixes #801.
Remove the JAX/JAXlib
<0.11bounds so a downstream project can select JAX 0.11. Refresh only these two packages inuv.lock, regenerateasv-constraints.txt, and add an Unreleased note. The lock selects 0.11.2 for Python >=3.12 and retains 0.10.2 for Python 3.11. No Python implementation, test, or CI configuration is changed.Validation
uv pip compilewith downstream constraintsjax>=0.11andjaxlib>=0.11fails against the base metadata and resolves against this metadata.uv run poe test: 3229 passed, 1 existing skip on Windows CPU in Python 3.12 and Python 3.13 with JAX/JAXlib 0.11.2. The source was imported from this checkout.uv run poe lint-ci: all 156 files pass.poe all-testsalso passed its 41 docstring examples before reaching the ASV environment failure described below.uv build --wheeland installation of that wheel with JAX/JAXlib 0.11.2 into a fresh Python 3.12 environment pass. The installed package runs a JIT exact-GP posterior mean example (NumPy 2.5.3, NumPyro 0.22.0, Equinox 0.13.8).uv run asv check -E existingpasses.Limits
The configured ASV virtualenv check is not a clean pass on Windows: it builds the untouched base commit 0415ef8, then its existing
{env_dir}/bin/pythoninstall command cannot find the Windows interpreter. The separate existing-environment check does not replace that result. I did not change this configuration.The full
poe docs-cinotebook build is still running; no full documentation success is claimed. No GPU/TPU, performance benchmark, reporter's private downstream project, or every older dependency combination was tested. A separate exploratory example returning a GaussianDistribution object through JIT loses itslocattribute on both JAX 0.10.2 and 0.11.2; returning the posterior mean array within JIT passes. This pre-existing behavior is not changed here.AI assistance
OpenAI Codex autonomously prepared and tested this contribution under the account owner's authorization. No independent human review or personal production use is claimed. No third-party implementation code was copied. Co-author credit is in the commit.
Documentation follow-up: the configured full
poe docs-cibuild completed with exit 1. Notebook execution reports the Windows IPython/watermark cell-a 'Thomas Pinder'failing withUsageError: unrecognized arguments: Pinder'. The same unchanged cell exists at base 0415ef8; a direct execution in this environment reproduces that parsing error independently of GPJax/JAX. This is a local baseline cell check, not a full baseline documentation build. Full documentation success is not claimed; no documentation warnings, workflows, or timeouts were disabled.