Added `sample_kind` to `ModelOutput`, without breaking changes by pfebrer · Pull Request #168 · metatensor/metatomic

pfebrer · 2026-02-26T21:58:27Z

This PR substitutes #165.

It also implements sample_kind, with the short-term goal of enabling per-pair targets, but in this case in a way that doesn't break backward compatibility (new code using old ModelOutput might fail, but old code using the new ModelOutput won't). This is achieved by keeping the possibility to pass per_atom as an argument, and setting/getting it as a property, even though from now on the only thing ModelOutput stores is sample_kind.

Thanks to @Luthaf for letting me know that one can set optional arguments in torch exportable classes :)

If you agree with merging this one I will finalize it with some more tests ✌️

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

⚙️ Download Python wheels for this pull-request (you can install these with pip)

📚 Download documentation for this pull-request

pfebrer · 2026-03-13T21:48:29Z

This is ready for a final look. I checked that one can take the conda lammps-metatomic and use this modified metatomic version without problems to run an MD with a pet-mad.pt exported using the last metatomic release. The results of the MD are exactly the same as when running without the modified version.

Luthaf · 2026-03-25T17:06:59Z

Do not merge before #186, we need to have at least one release of metatomic-ase that works with an already released version of metatomic-torch

pfebrer · 2026-04-17T11:40:48Z

@HaoZeke I don't understand your change, we wanted it to work (not make engines fail) but deprecate it, issuing a warning so that the engines get changed before it gets removed. What's the problem with issuing a warning, it does not make the code fail no?

HaoZeke · 2026-04-17T12:05:02Z

@HaoZeke I don't understand your change, we wanted it to work (not make engines fail) but deprecate it, issuing a warning so that the engines get changed before it gets removed. What's the problem with issuing a warning, it does not make the code fail no?

yeah that's true, but there's also the PR description where I thought we don't even want deprecation warnings, but OK, rebasing on #199 and then extending it

pfebrer · 2026-04-17T12:08:40Z

Aaah ok, noo the goal was to keep it compatible until metatomic introduces compatibility breaks (when multiple backends are supported probably), and at that point just remove per_atom

…lity Co-Authored-By: Guillaume Fraux <guillaume.fraux@epfl.ch> Co-Authored-by: Rohit Goswami <rgoswami@ieee.org>

pfebrer · 2026-04-21T10:37:39Z

Yessss, do you know approx when this will be in a realease?

Luthaf · 2026-04-21T12:10:58Z

End of this week/next week is the tentative deadline, but it will depend on how many other bugs I find in the mean time =)

pfebrer force-pushed the sample_kind branch from c73b0ad to d751945 Compare February 26, 2026 22:10

Luthaf reviewed Feb 27, 2026

View reviewed changes

Comment thread metatomic-torch/include/metatomic/torch/model.hpp Outdated

Comment thread metatomic-torch/src/model.cpp Outdated

Comment thread metatomic-torch/src/outputs.cpp Outdated

pfebrer force-pushed the sample_kind branch 3 times, most recently from 4b3c32d to 87dc377 Compare March 13, 2026 21:37

Luthaf reviewed Mar 16, 2026

View reviewed changes

pfebrer mentioned this pull request Mar 17, 2026

Replace per_atom with sample_kind everywhere. metatensor/metatrain#1075

Open

Luthaf force-pushed the sample_kind branch 4 times, most recently from dff557b to ea5e7fd Compare March 25, 2026 16:50

Luthaf force-pushed the sample_kind branch 3 times, most recently from 9353d37 to 4df02e3 Compare April 8, 2026 08:42

Luthaf force-pushed the sample_kind branch from 4df02e3 to fca4e05 Compare April 15, 2026 11:58

HaoZeke force-pushed the sample_kind branch 2 times, most recently from a24c33e to b387342 Compare April 17, 2026 10:45

HaoZeke requested a review from Luthaf April 17, 2026 10:56

HaoZeke force-pushed the sample_kind branch from b387342 to 67cf79c Compare April 17, 2026 10:57

HaoZeke force-pushed the sample_kind branch from 67cf79c to 1ab3831 Compare April 17, 2026 12:24

Luthaf reviewed Apr 17, 2026

View reviewed changes

Comment thread python/metatomic_ase/tests/conftest.py Outdated

Comment thread python/metatomic_ase/pyproject.toml Outdated

Luthaf force-pushed the sample_kind branch from 6f420ed to 4940a0b Compare April 21, 2026 08:54

Change per_atom into sample_kind while keeping backward compatibi…

0389f6b

…lity Co-Authored-By: Guillaume Fraux <guillaume.fraux@epfl.ch> Co-Authored-by: Rohit Goswami <rgoswami@ieee.org>

Luthaf force-pushed the sample_kind branch from 4940a0b to 0389f6b Compare April 21, 2026 09:19

Luthaf merged commit 256a66e into main Apr 21, 2026
10 checks passed

Luthaf deleted the sample_kind branch April 21, 2026 09:46

Luthaf mentioned this pull request Apr 21, 2026

Make sure deprecation warnings don't reach users that can not act on them #205

Merged

4 tasks

Conversation

pfebrer commented Feb 26, 2026 • edited by github-actions Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

pfebrer commented Mar 13, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Luthaf commented Mar 25, 2026

Uh oh!

pfebrer commented Apr 17, 2026

Uh oh!

HaoZeke commented Apr 17, 2026

Uh oh!

pfebrer commented Apr 17, 2026

Uh oh!

Uh oh!

Uh oh!

Uh oh!

pfebrer commented Apr 21, 2026

Uh oh!

Luthaf commented Apr 21, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

pfebrer commented Feb 26, 2026 •

edited by github-actions Bot

Loading