DEPLOY A MODEL

Write a deployment spec

Set MODEL_ID to the model you want to serve, then paste the block:

MODEL_ID=1

cat > deployment.json <<EOF
{
  "displayName": "support-classifier",
  "modelSpec": {
    "type": "custom",
    "modelId": $MODEL_ID,
    "autoscaling": { "minReplicas": 0, "maxReplicas": 1, "scaleDownIdleSeconds": 60 }
  }
}
EOF

Create the deployment

oumi-cli deployments create --input-json-file deployment.json

Returns the deployment ID.

Retrieve deployment info

oumi-cli deployments endpoint <DEPLOYMENT_ID>

Prints the inference URL, the model path, and a runnable curl snippet.

Evaluate a model

Score the model you deployed.

Deployments reference

Every flag on every deployment command, including inference log export.

⌘I

Getting started

Basics

Workflows

Command reference

DEPLOY A MODEL

Evaluate a model

Deployments reference