A lightweight wrapper for PyTorch that provides a simple declarative API for context switching between devices, distributed modes, mixed-precision, and PyTorch extensions.
It would be nice to see examples/support for how to use stoke with GAN training
Motivation
It is unclear how to train a GAN using stoke and if 2 stoke objects would need to be initialized (for generator and discriminator) or if it could be adapted to fit into a main stoke object
Proposal
Add necessary example for training a simple GAN or extend support if it cannot work out of the box
Feature
It would be nice to see examples/support for how to use stoke with GAN training
Motivation
It is unclear how to train a GAN using stoke and if 2 stoke objects would need to be initialized (for generator and discriminator) or if it could be adapted to fit into a main stoke object
Proposal
Add necessary example for training a simple GAN or extend support if it cannot work out of the box