History

Ning e8a6feb229 Restructure samples (#1710 ) * restructure samples * update apiserver and sample test for the new location		2019-08-01 17:31:37 -07:00
..
.s2i	Restructure samples (#1710 )	2019-08-01 17:31:37 -07:00
model_files	Restructure samples (#1710 )	2019-08-01 17:31:37 -07:00
README.md	Restructure samples (#1710 )	2019-08-01 17:31:37 -07:00
Serving.py	Restructure samples (#1710 )	2019-08-01 17:31:37 -07:00
requirements.txt	Restructure samples (#1710 )	2019-08-01 17:31:37 -07:00

README.md

PyTorch Gender Classification Seldon Serving container

Wrap the Runtime Scorer

You can skip this step if you are happy to use the already packaged image aipipeline/seldon-pytorch:0.1 from DockerHub.

The runtime MNIST scorer is contained within a standalone python class Serving.py. This needs to be packaged in a Docker container to run within Seldon. For this we use Redhat's Source-to-image.

Install S2I
From this seldon-pytorch-serving-image folder, run the following s2i build. You will need to change aipipeline to your DockerHub repo.

s2i build . seldonio/seldon-core-s2i-python2:0.4 aipipeline/seldon-pytorch:0.1

Push image to DockerHub or your Docker registry that's accessible from the KubeFlow Pipeline cluster.

docker push aipipeline/seldon-pytorch:0.1