Curiously, the master branch (updated about two months ago) has had the builds all passed, while the exact same code now apparently cannot pass the builds. The test failures are due to two estimator checks (autoencoder and MLP classifier) not passed. The environments (packages, etc.) should be identical between two months and this week, but maybe there's something we've overlooked.
I came across this while investigating why the builds didn't pass at #101.
Curiously, the master branch (updated about two months ago) has had the builds all passed, while the exact same code now apparently cannot pass the builds. The test failures are due to two estimator checks (autoencoder and MLP classifier) not passed. The environments (packages, etc.) should be identical between two months and this week, but maybe there's something we've overlooked.
I came across this while investigating why the builds didn't pass at #101.