Support numpy 2, upgrade tests to support torch 2.6 #221
Conversation
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
|
Pytables seems to be the issue (although the issue is closed PyTables/PyTables#1172) |
|
ah, pytables is a pain ... we should consider moving to |
|
Related: h5py/h5py#2420 (but also closed) |
stes
force-pushed
the
stes/upgrade-numpy
branch
from
1ccc466 to
a012229
Compare
|
Ok, good call to upgrade the test suite to pytorch 2.6 (latest), causes the following to fail 
I will push a fix |
stes
force-pushed
the
stes/upgrade-numpy
branch
from
8f30768 to
f082cae
Compare
stes
changed the title
2 tasks
CeliaBenquet
pushed a commit
to CeliaBenquet/CEBRA
that referenced
this pull request
Apr 23, 2025
…trolLab#221) * Drop numpy constraint * Implement workaround for pytables * better error message * pin numpy only for python 3.9 * update dependencies * Upgrade torch version * Fix based on python version * Add support for torch.load with weights_only=True * Implement safe loading for torch models starting in torch 2.6 * Fix windows specs * fix docstring * Revert changes to loading logic
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Numpy was pinned to
1.xbefore. Checking if an upgrade to2.xis possible.Background: The dependency was pinned in #172 due to issues with sklearn and docs building. Given the updates we made to fix new sklearn versions (#205 etc.), let's explore if we can drop the constraint by now.
New insights: PyTables is the remaining root cause preventing us from upgrading fully. This is only the issue for the
pd.read_hdffunction. The proposed fix is to attempt loading withpandasviapytables, but if this fails we can manually load usingh5py, and then construct thepandasdataframe from there.There are new libraries like
polarscoming up, but it seems like these do not have hdf support yet (https://docs.pola.rs/api/python/stable/reference/io.html).Current status: