lakshayg

The binary you are using was compiled with support for `fma avx avx2 sse4_1 sse4_2`. Your flags output does not contain `fma avx avx2` which means that your machine is...

Cool, can you please create a `contrib` directory and place the file in there?

1. TensorFlow is optimized for your CPU by default (`-march=native`) unless you specify different flags while compiling. If you want to enable only specific features, you can add flags like...

Are you saying that the compiled tensorboard does not work?

How about encrypting the user's private key or the whole of felony.json with a master password which the user must provide every time?

@pytorchbot label "topic: not user facing"

> > @lakshayg: @pytorchbot label "topic: not user facing" > > @nWEIdia: I would argue this is `user facing` , ... I have added the "release notes: cuda" label. Is...

@nWEIdia I have added unit tests for this PR. Let me know if you think of more cases that I should test.

@atalman This check I have written is based on this page: https://developer.nvidia.com/cuda-gpus. The rule I have followed is that higher CC device can run lower CC code in the same...

@nWEIdia I believe @ptrblck wanted to review it as well. He wanted to think through the case when someone is building pytorch with custom arch list.