- Docker
- GO (>= 1.11) for go API
- Python for python API
-
Window/MacOS: install Docker from DockerHub. (need to register new account if you are newbie)
-
linux: install Docker
$ git clone https://github.com/Alwaysproblem/MLserving-tutorial
$ cd MLserving-tutorial/TFserving/- here is no batch configuration only for demonstration
$ docker run --rm -p 8500:8500 -p 8501:8501 -v `pwd`:/models -it tensorflow/serving --model_config_file=/models/config/versionlabels.config --model_config_file_poll_wait_seconds=60 --allow_version_labels_for_unavailable_models- test for network
$ curl -d '{"instances": [[1.0, 2.0]]}' -X POST http://localhost:8501/v1/models/Toy_double:predict