创建docker镜像

在serving目录，使用tensorflow serving提供的dockerfile

$ docker build --pull -t $USER/tensorflow-serving-devel -f tensorflow_serving/tools/docker/Dockerfile.devel .
$ docker run --name=inception_container -it $USER/tensorflow-serving-devel

在运行的容器中，clone，configure，build所需要的tensorflow serving示例

root@c97d8e820ced:/# git clone --recurse-submodules https://github.com/tensorflow/serving
root@c97d8e820ced:/# cd serving/tensorflow
root@c97d8e820ced:/serving/tensorflow# ./configure
root@c97d8e820ced:/serving# cd ..
root@c97d8e820ced:/serving# bazel build -c opt tensorflow_serving/example/...

构建ModelServer

root@c97d8e820ced:/serving# bazel build -c opt tensorflow_serving/model_servers:tensorflow_model_server

导出Inception model

下载inception的checkpoint

root@c97d8e820ced:/serving# curl -O http://download.tensorflow.org/models/image/imagenet/inception-v3-2016-03-01.tar.gz
root@c97d8e820ced:/serving# tar xzf inception-v3-2016-03-01.tar.gz
root@c97d8e820ced:/serving# ls inception-v3
README.txt  checkpoint  model.ckpt-157585

导入checkpoint，并导出模型

root@c97d8e820ced:/serving# bazel-bin/tensorflow_serving/example/inception_saved_model --checkpoint_dir=inception-v3 --output_dir=/tmp/inception-export
Successfully loaded model from inception-v3/model.ckpt-157585 at step=157585.
Successfully exported model to /tmp/inception-export
root@c97d8e820ced:/serving# ls /tmp/inception-export
1

commit镜像

$ docker commit inception_container $USER/inception_serving
$ docker stop inception_container

运行镜像

$ docker run -it $USER/inception_serving

启动server

root@f07eec53fd95:/# cd serving
root@f07eec53fd95:/serving# bazel-bin/tensorflow_serving/model_servers/tensorflow_model_server --port=9000 --model_name=inception --model_base_path=/tmp/inception-export &> inception_log &
[1] 45

查询服务

root@f07eec53fd95:/serving# bazel-bin/tensorflow_serving/example/inception_client --server=localhost:9000 --image=/path/to/my_cat_image.jpg