Bugfix for #6143, flakey K-means test in GP submodule

[bwengals]

closes #6143.

One of the tests of the GP model tests that the kmeans method for locating inducing points works correctly. It uses an example that should be easy for K-means, a 1d normal mixture model, with mu at -5 and the other at 5, both with sigma=1. Occasionally the tests give an answer that is close enough, like [-4.98, 5.1], but doesn't pass the tests. This PR uses a 5% relative tolerance instead.

Checklist

• Explain important implementation details 👆
• Make sure that the pre-commit linting/style checks pass.
• Link relevant issues (preferably in nice commit messages)
• Are the changes covered by tests and docstrings?
• Fill out the short summary sections 👇

Major / Breaking Changes

• NA

Bugfixes / New features

• fix for Flaky tests/gp/test_util.py::TestKmeansInducing::test_kmeans #6143

Docs / Maintenance

• NA

[codecov]

Codecov Report

Merging #6157 (5e7d348) into main (b12c189) will increase coverage by 5.07%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #6157      +/-   ##
==========================================
+ Coverage   87.94%   93.02%   +5.07%     
==========================================
  Files          91       91              
  Lines       20707    20707              
==========================================
+ Hits        18211    19262    +1051     
+ Misses       2496     1445    -1051     

Impacted Files	Coverage Δ
pymc/tests/gp/test_util.py	100.00% <100.00%> (ø)
pymc/smc/smc.py	97.24% <0.00%> (+0.78%)	⬆️
pymc/step_methods/hmc/base_hmc.py	90.55% <0.00%> (+0.78%)	⬆️
pymc/parallel_sampling.py	86.79% <0.00%> (+0.99%)	⬆️
pymc/distributions/censored.py	100.00% <0.00%> (+7.50%)	⬆️
pymc/distributions/simulator.py	87.23% <0.00%> (+53.19%)	⬆️
pymc/distributions/bound.py	100.00% <0.00%> (+54.36%)	⬆️
pymc/distributions/truncated.py	99.30% <0.00%> (+64.33%)	⬆️
pymc/sampling_jax.py	97.18% <0.00%> (+97.18%)	⬆️
pymc/tests/distributions/test_simulator.py	99.50% <0.00%> (+99.50%)	⬆️
... and 3 more

[bwengals]

thanks for flagging this Ricardo