Skip to content

Update to metatensor-learn v0.6.0 - #1240

Draft
Luthaf wants to merge 3 commits into
mainfrom
mts-learn-0.6
Draft

Update to metatensor-learn v0.6.0#1240
Luthaf wants to merge 3 commits into
mainfrom
mts-learn-0.6

Conversation

@Luthaf

@Luthaf Luthaf commented Aug 3, 2026

Copy link
Copy Markdown
Member

This will later allow to update the serialization format and thus make it mmap-compatible.

This does not work yet, and requires #1238 to be merged first (trying out the new github feature about stacked PR here)

Still TBD:

  • fix issue with model.to() on an empty dict
  • Update other uses of metatensor values as attribute in the code

Contributor (creator of pull-request) checklist

  • Tests updated (for new features and bugfixes)?
  • Documentation updated (for new features)?
  • Issue referenced (for PRs that solve an issue)?

Maintainer/Reviewer checklist

  • CHANGELOG updated with public API or any other important changes?
  • GPU tests passed (maintainer comment: "cscs-ci run")?

📚 Documentation preview 📚: https://metatrain--1240.org.readthedocs.build/en/1240/

Base automatically changed from python-3.14 to main August 4, 2026 11:15
@Luthaf
Luthaf force-pushed the mts-learn-0.6 branch 4 times, most recently from f4635be to 5fe38ee Compare August 6, 2026 12:45
Luthaf added 2 commits August 6, 2026 14:46
Otherwise there can be a situation where different process from
xdist check the locks separately and aquire it, but then are interupted
and leave dangling locks, preventing future tests to from running
This simplify the forward that no longer needs to move data to a
different dtype. Instead the model is moved once in the trainer
and then stays on the same device
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant