To train a deployable object detection model, run:
cacli train --frameworkv 1.15 --script https://github.com/cloud-annotations/training/releases/download/v1.2.30/deployable.zip
Once training has finished, run the following to store the model:
ibmcloud ml store training-runs <CACLI-MODEL-ID>
You should get the following output:
Starting to store the training-run '<CACLI-MODEL-ID>' ...
OK
Model store successful. Model-ID is '18510cf7-2a9b-4677-9280-b4604e2e10cb'.
Use the returned information to run the deploy
command:
ibmcloud ml deploy 18510cf7-2a9b-4677-9280-b4604e2e10cb <CHOOSE-A-DEPLOYMENT-NAME>
You should get a response back with a scoring endpoint
Clone the repo:
git clone https://github.com/cloud-annotations/wml-api-demo
cd wml-api-demo
Create a .env
file with the following information:
APIKEY=<YOUR-WML-APIKEY>
SCORING_ENDPOINT=<YOUR-SCORING_ENDPOINT>
Install the dependencies:
npm install
Run the app:
npm start