- Build docker image:
docker build --pull -t $USER/tensorflow-serving-devel -f Dockerfile.devel .
- Run the docker container:
docker run --name=tf_container -it $USER/tensorflow-serving-devel
Copy the models into the container
cd <path to project>
docker cp ./export tf_container:/serving
- Start it again with
docker start -i tf_container
- Start serving (after you enter docker container)
tensorflow_model_server --port=9000 --model_name=MODELNAME --model_base_path=EXPORTPATH &> logs &
- Get container IP address
docker network inspect bridge | grep IPv4Address
- Test your client
cd <path to project>
python MODEL_client.py --server=<IP ADDRESS>:9000 ...