Always getting modulenotfound for deployment using yaml file

Ray Libraries (Data, Train, Tune, Serve)
1

Hello,

I have one application with single model defined in a python file let’s say inference:app where app I defined as a variable Model.bind() , everytime I try to deploy a python file using serve deploy config.yaml and when I check status it always throwing module not found error the model I want to deploy. My question is how to deploy using config.yaml a local python file which contains the model deployment code ? I am sharing my yaml file little bit later.

I tried following this documentation but it doesn’t worked.
https://docs.ray.io/en/latest/serve/api/doc/ray.serve.Application.html

My yaml looks like this :

proxy_location: ...

http_options: 
  host: ...
  port: ...
  request_timeout_s: ...
  keep_alive_timeout_s: ...

grpc_options:
  port: ...
  grpc_servicer_functions: ...

logging_config:
  log_level: ...
  logs_dir: ...
  encoding: ...
  enable_access_log: ...

applications:
- name: ...
  route_prefix: ...
  import_path: inference:app
  runtime_env: ... 
  deployments:
  - name: ...
    num_replicas: ...
    ...

My another question is this why it always replace the default application deployed to Ray serve , my ray version is : 2.9.3