pytorch / multipy

torch::deploy (multipy for non-torch uses) is a system that lets you get around the GIL problem by running multiple Python interpreters in a single C++ process.
Other
174 stars 35 forks source link

Broken nvidia-docker + python3.8 unittests #233

Open PaliC opened 2 years ago

PaliC commented 2 years ago

Stack from ghstack (oldest at bottom):

This PR should highlight some odd flakiness with python3.8 + nvidia-docker. This somehow leads to autograd failing in torch::deploy?