New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Link statically against TensorFlow C library #83

Open
lastzero opened this Issue Dec 19, 2018 · 2 comments

Comments

2 participants
@lastzero
Copy link
Member

lastzero commented Dec 19, 2018

As a user, I want to download a single binary for PhotoPrism so that I can easily use it on any computer without complicated installation.

Key to this endeavour is that we stick with TensorFlow as our only C library dependency and link it statically into our binary (maybe dlib too, another ML library). While it is currently not available in a static version from Google, we can build it ourselves or wait until next year when Bazel has a rule to build a fully static library.

AFAIK, a static version of TensorFlow can be built with cmake (there are howtos out there for various platforms): https://github.com/tensorflow/tensorflow/tree/master/tensorflow/contrib/cmake

Acceptance Criteria:

  • On Linux, make build MUST create a photoprism binary that is statically linked against TensorFlow (no shared libraries needed anymore)
  • Support for other operating systems and processors MAY be implemented
  • Support for other operating systems and processors SHOULD be documented (how to do it)
  • Support for CPU extensions, such as SSE4.1, SSE4.2, AVX, and AVX2 SHOULD be enabled

@lastzero lastzero added this to the MVP milestone Dec 19, 2018

@lastzero lastzero added the funded label Dec 20, 2018

@IssuehuntBot

This comment has been minimized.

Copy link

IssuehuntBot commented Dec 20, 2018

@lastzero has funded $40.00 to this issue. See it on IssueHunt

@IssuehuntBot

This comment has been minimized.

Copy link

IssuehuntBot commented Dec 21, 2018

@IssueHuntFest has funded $20.00 to this issue. See it on IssueHunt

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment