Computer Vision Demo

This repo helps to demonstrate the use of computer vision, containers, and openshift

Quickstart

Build model server via source

s2i strategy: source

APP_NAME=model-serving-cv

oc new-app \
  https://github.com/redhat-na-ssa/demo-computer-vision.git#peer-review \
  --name ${APP_NAME} \
  --strategy source \
  --context-dir /components/model

oc label all --all app=demo-computer-vision

Build model server via Dockerfile

s2i strategy: docker

APP_NAME=model-serving-cv-dockerfile
oc new-app \
  https://github.com/redhat-na-ssa/demo-computer-vision.git#peer-review \
  --label app=demo-computer-vision \
  --name ${APP_NAME} \
  --strategy docker \
  --context-dir /components/model

Expose API / model server - Route

oc expose service \
  ${APP_NAME} \
  --label app=demo-computer-vision \
  --port 8080 \
  --overrides='{"spec":{"tls":{"termination":"edge"}}}'

Build camera capture via Dockerfile

s2i strategy: docker

APP_NAME=camera-capture-cv
oc new-app \
  https://github.com/redhat-na-ssa/demo-computer-vision.git#peer-review \
  --label app=demo-computer-vision \
  --name ${APP_NAME} \
  --strategy docker \
  --context-dir /components/camera

Setup Liveness Probe

oc set probe deploy/${APP_NAME} \
  --liveness \
  --get-url=http://:8080/healthz

TODO

• tell story
• address dependencies