[tmva][sofie] Require onnx!=1.19.0 and keras<3.5 for tests#20575
Merged
guitargeek merged 2 commits intoroot-project:masterfrom Nov 30, 2025
Merged
[tmva][sofie] Require onnx!=1.19.0 and keras<3.5 for tests#20575guitargeek merged 2 commits intoroot-project:masterfrom
onnx!=1.19.0 and keras<3.5 for tests#20575guitargeek merged 2 commits intoroot-project:masterfrom
Conversation
1 task
guitargeek
force-pushed
the
onnx_version_check
branch
from
a1f73a4 to
f2bc7b7
Compare
guitargeek
changed the title
onnx>=1.19.1onnx>=1.19.1 for tests
guitargeek
force-pushed
the
onnx_version_check
branch
3 times, most recently
from
020734a to
dfe76cc
Compare
guitargeek
changed the title
onnx>=1.19.1 for testsonnx!=1.19.0 for tests
guitargeek
force-pushed
the
onnx_version_check
branch
2 times, most recently
from
7332842 to
9a267f9
Compare
guitargeek
changed the title
onnx!=1.19.0 for testsonnx!=1.19.0 and keras<3.5 for tests
This is a followup to 10f28b6, where I used a random keras version in a check in order to make the CI configuration at the time pass. In fact, `keras=>3.5` also didn't work, as we see now after updating the macOS runners from Keras 2 to Keras 3. It could actually turn out that `keras>=3` is not supported at all, so we should be prepared to lower the maximum supported Keras version even further. But for now we don't know, as no platform has `keras>=3.0&&keras<3.5` installed.
The Python package onnx 1.19.0 has a bug that makes this version unusable: onnx/onnx#7249 In that case, we have to disable the "TestSofieModels" and "TestRModelParserPyTorch" tests, which import onnx indirectly via `torch.onnx`. We should also consider to require `onnx!=1.19.1` in our `requirements.txt` in the future, so our users don't face similar trouble from exporting PyTorch models to onnx. But this should only be done once we are sure that it can also be installed on macOS without breaking something else. Closes root-project#20571.
guitargeek
force-pushed
the
onnx_version_check
branch
from
9a267f9 to
c6c851a
Compare
vepadulano
approved these changes
Nov 30, 2025
Member
vepadulano
left a comment
vepadulano
left a comment
There was a problem hiding this comment.
Thank you so much! It's nice to see the builds green :)
Test Results 22 files 22 suites 3d 16h 43m 43s ⏱️ Results for commit c6c851a. |
This was referenced Dec 1, 2025
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.
The Python package onnx 1.19.0 has a bug that makes this version unusable: onnx/onnx#7249. In that case, we have to disable the "TestSofieModels" and "TestRModelParserPyTorch" tests, which import onnx indirectly via
torch.onnx.We should also consider to require
onnx!=1.19.1in ourrequirements.txtin the future, so our users don't face similar trouble from exporting PyTorch models to onnx. But this should only be done once we are sure that it can also be installed on macOS without breaking something else.Closes #20571.
There is also a followup to guitargeek@10f28b6, where I used a random keras version in a check in order to make the CI configuration at the time pass. In fact,
keras=>3.5also didn't work, as we see now after updating the macOS runners from Keras 2 to Keras 3.It could actually turn out that
keras>=3is not supported at all, so we should be prepared to lower the maximum supported Keras version even further. But for now we don't know, as no platform haskeras>=3.0&&keras<3.5installed.