Running In Production

Although you can use in order start up your service, it is only recommended for local development.

We encourage to use gunicron for production use.

gunicorn with gevent is battle-tested works well for most use-cases:

Gunicorn Installation

$ pip install gunicorn[gevent]


Serve your application:

# example

from mlserving import ServingApp
# other imports ...

app = ServingApp()

Run: $ gunicorn -b -k gevent -w 4 app:app

Read more here