search

brendanhasz / probflow

A Python package for building Bayesian models with TensorFlow or PyTorch
http://probflow.readthedocs.io
MIT License
171 stars 17 forks source link

Contribution Guide #5

Closed jejjohnson closed 5 years ago

jejjohnson commented 5 years ago

Hello!

First of all, I really like this package you've created! It's a really nice stepping stone into the world of Bayesian NN without the need for a lot of boilerplate code.

But, I was wondering if you had a priority list potential contributions? Also perhaps a schematic of what type of things should be added or what things are out of the scope of this package?

Thanks, Emmanuel

jejjohnson commented 5 years ago

Nevermind, I found your list in the backlog section!

Closing.

brendanhasz commented 5 years ago

Hi! Thanks, glad you like the package! Yep, the priority list is in the backlog section, though I could add a list of things which are out-of-scope, or maybe a contributing guide, or a schematic, if you'd still find any of those useful.

Cheers,

Brendan

jejjohnson commented 5 years ago

I think it would find it useful for potential devs to at least know what's out of the scope.

Thanks again!

brendanhasz commented 5 years ago

Added an out-of-scope section to the backlog page, thanks for the idea!

Unfortunately, I included Gaussian Processes in that list (and it appears you've done a lot of work with GPs - so, sorry! haha) since I found it a bit hard to figure out how to fit nonparametric models into the ProbFlow framework (e.g., the probflow.model.predict interface would have to change, and all the methods which depend on it, since you'd need both the x and y training data as well as the x test data). For more info on that see https://github.com/brendanhasz/probflow/issues/7