Package: scikit-learn / 1.2.1+dfsg-1
Metadata
Package | Version | Patches format |
---|---|---|
scikit-learn | 1.2.1+dfsg-1 | 3.0 (quilt) |
Patch series
view the series filePatch | File delta | Description |
---|---|---|
sphinx doc no download.patch | (download) |
examples/applications/plot_out_of_core_classification.py |
3 2 + 1 - 0 ! |
prevent dataset download during doc build. guard with debian specific environment variable to avoid breaking the functionality. |
relax_test_tolerance.patch | (download) |
sklearn/neural_network/tests/test_mlp.py |
2 1 + 1 - 0 ! |
relax test tolerance - this patch is needed at least for i386 (not for amd64) > assert_allclose(rbm_64.components_, rbm_32.components_, rtol=1e-03, atol=0) E AssertionError: E Not equal to tolerance rtol=0.001, atol=0 E E Mismatched elements: 1 / 1024 (0.0977%) |
Disable BinderHub links.patch | (download) |
doc/conf.py |
8 0 + 8 - 0 ! |
disable binderhub links |
parallel cythonize.patch | (download) |
sklearn/_build_utils/__init__.py |
6 1 + 5 - 0 ! |
get cythonize parallelization from environment Instead of hard-coding to the CPU count, get the number of jobs for the JOBS env var (eg: set in debian/rules), defaulting to 1 if it is not set. |
no download intersphinx mapping.patch | (download) |
doc/conf.py |
13 5 + 8 - 0 ! |
no-downloading intersphinx mapping =================================================================== |
remove google analytics.patch | (download) |
doc/conf.py |
2 1 + 1 - 0 ! |
disable google analytics referrals in the -doc if we were to provide some information from just reading documentation we had to disclose and seek approval from the user before doing that. Thanks Ambrose Andrews for the analysis and report |
git revision cmd.patch | (download) |
doc/sphinxext/github_link.py |
4 2 + 2 - 0 ! |
supply github revision for built documentation This is used for the [source] links in the documentation. |
no buttons js.patch | (download) |
doc/developers/contributing.rst |
1 0 + 1 - 0 ! |
disable javascript github buttons |
Use local MathJax.patch | (download) |
doc/conf.py |
2 1 + 1 - 0 ! |
use local mathjax |
align create_memmap_backed_data.patch | (download) |
sklearn/utils/_testing.py |
2 1 + 1 - 0 ! |
align create_memmap_backed_data() by default This avoids a bus error on armhf and should avoid slow downs on other architectures, thanks Seth Arnold Bug-Ubuntu: https://bugs.launchpad.net/ubuntu/+source/scikit-learn/+bug/1991621 |
fix_clean.patch | (download) |
setup.py |
3 3 + 0 - 0 ! |
provide all method to clean target |