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

Integrate Tensorflow via JavaCPP as a proper service. #209

Closed
supertick opened this Issue Jun 2, 2018 · 3 comments

Comments

3 participants
@supertick
Contributor

supertick commented Jun 2, 2018

No description provided.

@supertick supertick created this issue from a note in Nixie Release (To Do) Jun 2, 2018

@kwatters

This comment has been minimized.

Contributor

kwatters commented Jun 5, 2018

There seem to be some windows dependency issues in the javacpp tensorflow bindings for tesseract 1.7. I think we should wait until the tensorflow 1.8 bindings are released. (perhaps we can/should help them with that release.)

@kwatters

This comment has been minimized.

Contributor

kwatters commented Jun 6, 2018

Turns out Tensorflow has maven java dependencies ! Initial tensorflow service created with dependencies. no actual implementation yet. but it's borged and ready to be added to.

Of note, Tensorflow's binaries only support windows/ linux / mac .. (as far as I can tell, no ARM support at this time.. so if we want to support tensorflow on the raspi or jetson, we'll need to re-visit this implementation.

@kwatters kwatters closed this Jun 6, 2018

Nixie Release automation moved this from To Do to Done Jun 6, 2018

@saudet

This comment has been minimized.

saudet commented Jun 16, 2018

If you guys can come up with linux-arm64 and linux-arm64-gpu builds with cross-compilation support for the JavaCPP Presets for TensorFlow, I'll be happy to maintain them and provide binaries:
https://github.com/bytedeco/javacpp-presets/tree/master/tensorflow

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