Skip to content

Handle TensorFlow like OpenCV#18185

Merged
pvizeli merged 2 commits intodevfrom
fix-tensorflow
Nov 4, 2018
Merged

Handle TensorFlow like OpenCV#18185
pvizeli merged 2 commits intodevfrom
fix-tensorflow

Conversation

@pvizeli
Copy link
Copy Markdown
Member

@pvizeli pvizeli commented Nov 4, 2018

Description:

That is the same problem as we have with OpenCV. The python wheels are not available for all platforms (amd64/i386/arm/aarch64) and OS.

With this change, the user needs self-install TensorFlow like OpenCV. Need also an update on Documentation.

@pvizeli pvizeli added this to the 0.82 milestone Nov 4, 2018
@pvizeli
Copy link
Copy Markdown
Member Author

pvizeli commented Nov 4, 2018

@hunterjm
Copy link
Copy Markdown
Member

hunterjm commented Nov 4, 2018

You are correct that the wheel is not available on all platforms, but if it is self compiled and pre-installed before this component runs, the component should use it, correct? Why not keep it here for those who do have it available to remove an additional step?

I suppose that removing it also allows a user to choose gpu or cpu if their setup supports it, so not totally against it.

@pvizeli
Copy link
Copy Markdown
Member Author

pvizeli commented Nov 4, 2018

Because it breaks all build system they have not the wheels. It's the same with dlib and OpenCV. wee need run a solution they bring the best compatibility.

@pvizeli pvizeli merged commit 27e159f into dev Nov 4, 2018
@ghost ghost removed the in progress label Nov 4, 2018
@balloob balloob deleted the fix-tensorflow branch November 4, 2018 14:19
balloob pushed a commit that referenced this pull request Nov 5, 2018
* Handle TensorFlow like OpenCV

* Update requirements_all.txt
@balloob balloob mentioned this pull request Nov 9, 2018
@home-assistant home-assistant locked and limited conversation to collaborators Feb 5, 2019
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants