1. A Cortex cluster (installation instructions)

  2. The Cortex CLI (installation instructions)


Let's deploy a classifier built using the famous iris data set!


$ mkdir iris && cd iris && touch cortex.yaml

Cortex requires a cortex.yaml file which defines a deployment resource. An api resource makes the model available as a live web service that can serve real-time predictions.

# cortex.yaml
- kind: deployment
name: iris
- kind: api
name: classifier
model: s3://cortex-examples/iris/tensorflow.zip
model_format: tensorflow

Cortex is able to read from any S3 bucket that you have access to.

Deploy the API

$ cortex deploy

You can get a summary of the status of resources using cortex get:

$ cortex get --watch

Test the API

Get the API's endpoint:

$ cortex get classifier

Use cURL to test the API:

$ curl -k -X POST -H "Content-Type: application/json" \
-d '{ "samples": [ { "sepal_length": 5.2, "sepal_width": 3.6, "petal_length": 1.4, "petal_width": 0.3 } ] }' \
<API endpoint>


Delete the deployment:

$ cortex delete iris