better / convoys

Implementation of statistical models to analyze time lagged conversions
https://better.engineering/convoys/
MIT License
254 stars 42 forks source link

Use tf.contrib.opt.ScipyOptimizerInterface #36

Closed erikbern closed 6 years ago

erikbern commented 6 years ago

Pretty sweet: https://www.tensorflow.org/api_docs/python/tf/contrib/opt/ScipyOptimizerInterface

Something with the Gamma convergence is still wacky – I suspect maybe tf.igamma might have numerical issues with its gradient. Ended up just using Powell for it, which doesn't require any gradient information. Full circle back to the pre-Tensorflow era.

coveralls commented 6 years ago

Coverage Status

Coverage decreased (-0.2%) to 91.534% when pulling c38fa120a662950e276ff141510d62c723ee289c on scipy-optimizer into ef01b4be0b419500299b4e3dffd4d1c9bed8c515 on master.

erikbern commented 6 years ago

This does better with Gamma on some real use cases I have, but sometimes gets stuck in a nan domain.

I'm also somewhat concerned this prevents using regularization strategies like dropout