Add ray to SageMaker Distribution v4.3#1254
Merged
Merged
Conversation
Add ray (ray-default) as an explicit top-level dependency for the cpu and gpu environments in v4.3. ray is currently present only transitively (via autogluon); this promotes it to a first-class, version-tracked package. Pinned >=2.53.0,<3.0.0 to stay within autogluon's supported ray range (autogluon 1.5.0 caps ray-default < 2.54) while allowing minor/ patch upgrades. Also adds a test Dockerfile validating ray import + local-mode init, registered in test_dockerfile_based_harness.py for both cpu and gpu.
HariniNarayanan
force-pushed
the
add-ray-package
branch
from
d30ffe6 to
f6184ad
Compare
TRNWWZ
approved these changes
Jul 7, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
Add
ray(ray-default) as an explicit top-level dependency of SageMaker Distribution v4.3, for both CPU and GPU images.Today
rayis only present transitively (pulled in byautogluon). This promotes it to a first-class, version-tracked package so it stays in the image on its own.How
conda-forge::ray-default[version='>=2.53.0,<3.0.0']tocpu.additional_packages_env.inandgpu.additional_packages_env.inunderbuild_artifacts/v4/v4.3/v4.3.0/.test/test_artifacts/v4/ray.test.Dockerfile(importsrayand runsray.init(local_mode=True)), registered in the CPU and GPU lists intest/test_dockerfile_based_harness.py.The floor is pinned to
>=2.53.0becauseautogluon 1.5.0(already in the image) requiresray-default <2.54, so2.53.0is the current compatible version; anything higher fails to solve.<3.0.0caps at the next major.Per CONTRIBUTING.md step 8, only the
additional_packages_env.infiles and test files are included;env.in/env.out/Dockerfileare regenerated by maintainers/CI.Testing
Built both images locally end-to-end via
python ./src/main.py build --target-patch-version=4.3.0:4.3.0-cpu);env.outresolvesray-default-2.53.0.4.3.0-gpu).The environment solves cleanly with
ray-defaultpromoted to a top-level dependency, co-existing withautogluonat the compatible2.53.0.