Make Predictions

You could just visit the endpoint (http://simple-server-4k2epq5lynxbaayn.192.168.71.93.modelz.live in this case) if you are using Gradio, streamlit or other web UI-based inference servers.

$ mdz deploy --image aikain/simplehttpserver:0.1 --name simple-server --port 80
Inference simple-server is created
$ mdz list
 NAME           ENDPOINT                                                          STATUS  INVOCATIONS  REPLICAS 
 simple-server  http://simple-server-4k2epq5lynxbaayn.192.168.71.93.modelz.live   Ready             2  1/1      
                http://192.168.71.93.modelz.live/inference/simple-server.default

You could also use the endpoint to send RESTful or websocket requests to the deployment.

$ curl http://simple-server-4k2epq5lynxbaayn.192.168.71.93.modelz.live

You could get the total number of requests in mdz list command.

$ mdz deploy --image aikain/simplehttpserver:0.1 --name simple-server --port 80
Inference simple-server is created
$ mdz list
 NAME           ENDPOINT                                                          STATUS  INVOCATIONS  REPLICAS 
 simple-server  http://simple-server-4k2epq5lynxbaayn.192.168.71.93.modelz.live   Ready             3  1/1      
                http://192.168.71.93.modelz.live/inference/simple-server.default

Create the First Deployment Autoscale the Deployment