[tmva][sofie] Remove Keras and PyTorch parses from ROOT repository - #22734
Open
guitargeek wants to merge 1 commit into
Open
[tmva][sofie] Remove Keras and PyTorch parses from ROOT repository#22734guitargeek wants to merge 1 commit into
guitargeek wants to merge 1 commit into
Conversation
guitargeek
force-pushed
the
sofie_keras_torch
branch
from
June 30, 2026 18:04
7aa73e9 to
28810e5
Compare
Test Results 23 files 23 suites 3d 12h 26m 51s ⏱️ For more details on these failures, see this check. Results for commit 64a153f. ♻️ This comment has been updated with latest results. |
guitargeek
force-pushed
the
sofie_keras_torch
branch
3 times, most recently
from
July 13, 2026 06:49
aff6bf7 to
b5be88e
Compare
guitargeek
force-pushed
the
sofie_keras_torch
branch
3 times, most recently
from
July 24, 2026 16:30
cb23ac9 to
b44f619
Compare
guitargeek
marked this pull request as ready for review
July 24, 2026 16:30
guitargeek
requested review from
bellenot,
couet,
dpiparo,
lmoneta and
vepadulano
as code owners
July 24, 2026 16:30
guitargeek
force-pushed
the
sofie_keras_torch
branch
from
July 24, 2026 16:48
b44f619 to
b8ac2ae
Compare
dpiparo
approved these changes
Jul 28, 2026
dpiparo
left a comment
dpiparo
left a comment
Member
There was a problem hiding this comment.
If lorenzo and sanjiban agree, LGTM.
guitargeek
force-pushed
the
sofie_keras_torch
branch
from
August 4, 2026 05:59
b8ac2ae to
d9f86be
Compare
As announced in the Q2 2026 Report and Planning meeting [1], the SOFIE distribution within ROOT will be slimmed down to keep only the core functionality of enabling ONNX model inference on the CPU with ROOT via ONNX parsing and C++ code emit. That means to keep the ONNX parser, and to remove the Keras and PyTorch parsers. these parsers relied on private implementation details of the Keras or PyTorch libraries, which change faster than appropriate for the stability standards of ROOT. We expect users to choose neural net architectures that are serializable by ONNX for production use. The Keras and PyTorch parsers remain valuable work, visible in ROOTs commit history and ready to be re-distributed by other parties, which would be very easy since this is just Python code. It would be trivial to ship this as a Python package that requires ROOT or another distribution of C++ SOFIE. [1] https://indico.cern.ch/event/1699702/ : slide 19
guitargeek
force-pushed
the
sofie_keras_torch
branch
from
August 14, 2026 07:22
d9f86be to
64a153f
Compare
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.
As announced in the Q2 2026 Report and Planning meeting [1], the SOFIE distribution within ROOT will be slimmed down to keep only the core functionality of enabling ONNX model inference on the CPU with ROOT via ONNX parsing and C++ code emit.
That means to keep the ONNX parser, and to remove the Keras and PyTorch parsers. these parsers relied on private implementation details of the Keras or PyTorch libraries, which change faster than appropriate for the stability standards of ROOT. We expect users to choose neural net architectures that are serializable by ONNX for production use.
The Keras and PyTorch parsers remain valuable work, visible in ROOTs commit history and ready to be re-distributed by other parties, which would be very easy since this is just Python code. It would be trivial to ship this as a Python package that requires ROOT or another distribution of C++ SOFIE.
[1] https://indico.cern.ch/event/1699702/ : slide 19
A demo for how it could be packaged outside of ROOT is given in: https://github.com/guitargeek/sofie-extra-parsers