CODEWITHSUNDEEP
tensorflow
rmg
rmg

Reputation: 1089

Tensorflow Serving basic example

I am working through the Tensorflow serving_basic example at:

https://tensorflow.github.io/serving/serving_basic

Setup

Following: https://tensorflow.github.io/serving/setup#prerequisites

Within a docker container based off of ubuntu:latest, I have installed:

bazel:

echo "deb [arch=amd64] http://storage.googleapis.com/bazel-apt stable jdk1.8" | sudo tee /etc/apt/sources.list.d/bazel.list
curl https://bazel.build/bazel-release.pub.gpg | sudo apt-key
sudo apt-get update && sudo apt-get install bazel
sudo apt-get upgrade bazel

grpcio:

pip install grpcio

all packages:

sudo apt-get update && sudo apt-get install -y         build-essential         curl         libcurl3-dev         git         libfreetype6-dev         libpng12-dev         libzmq3-dev         pkg-config         python-dev         python-numpy         python-pip         software-properties-common         swig         zip         zlib1g-dev

tensorflow serving:

git clone --recurse-submodules https://github.com/tensorflow/serving
cd serving
cd tensorflow
./configure
cd ..

I've built the source with bazel and all tests ran successfully:

bazel build tensorflow_serving/...
bazel test tensorflow_serving/...

I can successfully export the mnist model with:

bazel-bin/tensorflow_serving/example/mnist_export /tmp/mnist_model

And I can serve the exported model with:

bazel-bin/tensorflow_serving/model_servers/tensorflow_model_server --port=9000 --model_name=mnist --model_base_path=/tmp/mnist_model/

The problem

When I test the server and try to connect a client to the model server with:

bazel-bin/tensorflow_serving/example/mnist_client --num_tests=1000 --server=localhost:9000

I see this output:

root@dc3ea7993fa9:~/serving# bazel-bin/tensorflow_serving/example/mnist_client --num_tests=2 --server=localhost:9000
Extracting /tmp/train-images-idx3-ubyte.gz
Extracting /tmp/train-labels-idx1-ubyte.gz
Extracting /tmp/t10k-images-idx3-ubyte.gz
Extracting /tmp/t10k-labels-idx1-ubyte.gz
AbortionError(code=StatusCode.NOT_FOUND, details="FeedInputs: unable to find feed output images")
AbortionError(code=StatusCode.NOT_FOUND, details="FeedInputs: unable to find feed output images")

Inference error rate is: 100.0%

Upvotes: 1

Views: 1317

Answers (2)

rmg
rmg

Reputation: 1089

I mentioned this on the tensorflow github, and the solution was to remove the original model that had been created. If you're running into this, run rm -rf /tmp/mnist_model and rebuild it

Upvotes: 0

Sudhakar
Sudhakar

Reputation: 43

The "--use_saved_model" model flag is set to default "true"; use the --use_saved_model=false when starting the server. This should work:

bazel-bin/tensorflow_serving/model_servers/tensorflow_model_server --use_saved_model=false --port=9000 --model_name=mnist --model_base_path=/tmp/mnist_model/

Upvotes: 1

Related Questions