deprecation warnings for tensorflow

[jaanli]

deprecation warnings for tensorflow

[mathematicalmichael]

could really benefit from some outlining of the dependencies.
nomen, torch, etc. and what you're assuming about the environment.
I got this pretty absurd stack trace, and setting TMPDIR as an environment variable doesn't seem to do the trick.
It would also have helped to say that there's an assumption of an NVIDIA GPU.

I feel like I spent an inordinate amount of time trying to get your code to run (after reading your variational autoencoder article), and much of this could have been prevented with a few lines of environmental context in the README. =(

[jaanli]

Thanks @mathematicalmichael — you are right that this code is a few years old and needs to be updated

[jaanli]

@mathematicalmichael can you take a look at b43325e please?

It should suppress most of the warnings. Is that better now? Any other ways to make it easier to understand? I'm not sure I'm using tf.datasets in the simplest way, but I think feed_dict's are more intuitive than tfrecords...

I also added an environment.yml file for anaconda with the latest tensorflow version.