search

tobegit3hub / simple_tensorflow_serving

Generic and easy-to-use serving service for machine learning models
https://stfs.readthedocs.io
Apache License 2.0
757 stars 195 forks source link

Unable to run local trained model #43

Closed MaybeShewill-CV closed 5 years ago

MaybeShewill-CV commented 5 years ago

I have convert one of my pretrained network weights into tensorflow saved model and use the check tools you supplied which shows nothing wrong. But when I run the image inference from the broswer I got following mistake. Could you please help me to locate which step may be wrong. Thanks a lot. Here is some information question ques

tobegit3hub commented 5 years ago

Thanks for reporting.

This page is for image model which requires specified model signature. Could you try with the codegen python client to test with your model?

MaybeShewill-CV commented 5 years ago

@tobegit3hub Sorry I have solved the problem. I forget to close this issue. By the way can I ask for another help that may has little relationship with this project. The problem is that when I run a server within a tensorflow serving docker container and run a grpc client outside of the container everything works fine without vpn connection but it will not work when I establish the vpn connection. During the two process the server address are both localhost:9000. Could you please help me with it if you've got spare time:)

tobegit3hub commented 5 years ago

This seems to be the issue of your VPN. I don't know how you setup it and use it, so I have no methods to address this. But you can try run docker container with --net=host so that it uses the host network and make sure the client can access the server.

MaybeShewill-CV commented 5 years ago

@tobegit3hub Thanks a lot:)