llsourcell / how-to-deploy-a-tensorflow-model-in-production

This is the code for the "How to Deploy a Tensorflow Model in Production" by Siraj Raval on YouTube

How-to-Deploy-a-Tensorflow-Model-in-Production

This is the code for the "How to Deploy a Tensorflow Model in Production" by Siraj Raval on YouTube

Overview

This is the code for this video on Youtube by Siraj Raval. We're going to use the Tensorflow Serving library to deploy an inception model in production.

Dependencies

All included in the iPython notebook. You just need docker

Usage

Run the notebook by running jupyter notebook in terminal in the main directory. All the instructions are in there.

The 2 attached files -- one is the client and the other is an example of how we train and save an simple MNIST model for Tensorflow Serving.

Credits

Credits go to Google. I've merely created a wrapper to get people started.

how-to-deploy-a-tensorflow-model-in-production's Issues

Please update this article

Please update this article. Would greatly appreciate it.

'inception_export' no such file or directory

bazel-bin/tensorflow_serving/example/inception_export --checkpoint_dir=inception-v3 --export_dir=inception_export
bash: ./bazel-bin/tensorflow_serving/example/inception_export: No such file or directory

Failure: bazel build -c opt tensorflow_serving/...

ERROR: /root/.cache/bazel/_bazel_root/f8d1071c69ea316497c31e40fe01608c/external/org_tensorflow/tensorflow/core/kernels/BUILD:3867:1: C++ compilation of rule '@org_tensorflow//tensorflow/core/kernels:training_ops' failed: gcc failed: error executing command /usr/bin/gcc -U_FORTIFY_SOURCE -fstack-protector -Wall -B/usr/bin -B/usr/bin -Wunused-but-set-parameter -Wno-free-nonheap-object -fno-omit-frame-pointer -g0 -O2 '-D_FORTIFY_SOURCE=1' -DNDEBUG ... (remaining 115 argument(s) skipped): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 4.
gcc: internal compiler error: Killed (program cc1plus)

What is the version of tensorflow_serving and tensorflow is used in this tutorial?

I failed to build the example, could your please specify the version of tensorflow-serving and tensorflow that you are using in this tutorial

If it is not a release version, please give me the commit SHA1 Id, thanks.

How to deploy my model?

I have a working directory of my model, i want to know how to integrate it?

ImportError for backports module when setting checkpoint_dir with inception model

When "saving the inception-v3 model as a checkpoint" (bazel-bin/tensorflow_serving/example/inception_export --checkpoint_dir=inception-v3 --export_dir=inception-export), I get an ImportError for the backports module. I tried using pip install backports but seems like the container doesn't have what I need:

Collecting backports Could not find a version that satisfies the requirement backports (from versions:) No matching distribution found for backports

I'm trying to find a solution for this error in OpenStack, but no luck so far. I found out the current version of the Dockerfile.devel is different from the one you used in the video. For example, now the first line reads : FROM ubuntu:16.04, instead of FROM ubuntu:14.04 like is your case. Not sure what else was changed and if that is the reason for my error. Help?

EDIT: pip install backports.weakref did the trick. I'm sorry about this unfortunate post.