Davide Nunes

@apiad in case you're interested I also made some builds for python 3.* [here](https://github.com/davidenunes/tensorflow-wheels) if you need something for CPU only I can probably make it happen. It takes me...

I recognize that this requires an overhaul to allow for other people to extend things more easily. I just hastily put it together for my own purposes back then 😅...

Can you give me a precise example of something to be documented and a description of the expected result just so I can reproduce the problem more easily?

Thank you, this helps :+1:, I'm a bit swamped at the time but I'll give it a look as soon as I can

Sorry for the delay, this was an easy fix but I've been a bit swamped with my thesis work It should be working in version 0.9.2

Before I try to build for it, does the official install fail on your system? and if so, do you have any idea why?

I would probably need to spin a VM for the build, since I'm using Arch but perhaps I can help you, I need more information though Does the official install...

OK thanks, I was asking because I'm on a skylake if I don't specify anything it would target my architecture it seems that **-march=skylake-avx512** should work for the [Xeon Gold...

That's why I asked if there was other problem preventing you from using the official install I thought it could be something OS-related. So everything works? I ended-up building it...