Skip to content
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

contribute: Add the API to set the session configuration when load the model in the JAVA #18143

Closed
raintung opened this issue Mar 31, 2018 · 2 comments
Labels
stat:contribution welcome Status - Contributions welcome type:feature Feature requests

Comments

@raintung
Copy link

System information
Have I written custom code (as opposed to using a stock example script provided in TensorFlow): NO
OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Ubuntu
TensorFlow installed from (source or binary): source
TensorFlow version (use command below): Master
Python version: 3.6
Bazel version (if compiling from source): N/A
GCC/Compiler version (if compiling from source): N/A
CUDA/cuDNN version: None
GPU model and memory: None
Exact command to reproduce: Run the snippet below.

In the Java, only can set the run options and can't modify the session configuration when load the model. Sometimes I want to set the session strategies, ex. Set the session threads, op threads number, and so on.

I write the code to add new interface in the JAVA, and also add the JNI corresponding method.
I have been submit the code in the github:
URL:
https://github.com/raintung/TensorflowPatch

I will use it in my company project, I think it should not conflict the Tensorflow license, I still contribute it.

@tensorflowbutler
Copy link
Member

Nagging Assignee @skye: It has been 14 days with no activity and this issue has an assignee. Please update the label and/or status accordingly.

@skye skye assigned asimshankar and unassigned skye Apr 18, 2018
@asimshankar
Copy link
Contributor

If I understand correctly, you're proposing that a variant of SavedModelBundle.load() be added that takes in the session configuration (e.g., a serialized ConfigProto protocol buffer)?

Sure, that sounds perfectly reasonable.
Marking this as contributions welcome, feel free to send in a pull request. And I would suggest that we don't add a new native method, simply modify the existing one in place.

@asimshankar asimshankar added stat:contribution welcome Status - Contributions welcome type:feature Feature requests labels Apr 18, 2018
@asimshankar asimshankar removed their assignment Apr 18, 2018
raintung added a commit to raintung/tensorflow that referenced this issue Apr 20, 2018
raintung added a commit to raintung/tensorflow that referenced this issue Apr 20, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
stat:contribution welcome Status - Contributions welcome type:feature Feature requests
Projects
None yet
Development

No branches or pull requests

4 participants