Contributor

Types of changes

Variable default values.

Motivation and context / Related issue

Fix small typo in the code, where the parameter stopThr in empirical_sinkhorn_divergence was useless because subcalls used stopThr=1e-9 instead of stopThr=stopThr.

Same for a similar instance occuring in sinkhorn_epsilon_scaling.

Documented in Issue #421 .

PR checklist

• I have read the CONTRIBUTING document.
• The documentation is up-to-date with the changes I made (check build artifacts).
• All tests passed, and additional code has been covered with new tests.
• I have added the PR and Issue fix to the RELEASES.md file.

Contributor Author

It seems the doc build fails because of an issue with plot_gromov_barycenter.py.

Unexpected failing examples:
/home/circleci/project/examples/gromov/plot_gromov_barycenter.py failed leaving traceback:
Traceback (most recent call last):
  File "/home/circleci/project/examples/gromov/plot_gromov_barycenter.py", line 113, in <module>
    xs = np.array([np.array(xs[0]), np.array(xs[1]),
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (4,) + inhomogeneous part.

source

Did I do something wrong? I don't understand what I've done that could affect this example.

Collaborator

I just launched the run for doc build in CI l'et us see what is happening. This is weird indeed.

Collaborator

ok this is a new bug that appeared recently I will look into it.

Thanks for the PR it looks perfect I will mere it as soon as I handled this weird new bug.

Collaborator

Thanks very much for the two PR @tlacombe it is appreciated