4 years ago

Category: Data

In order to deploy a model at an endpoint, you will first need to select a web-framework to use to create that endpoint. For Python and R, the choice is pretty obvious, and it can be rather easy to find a proper web-framework that will meet your needs.

This will allow you to use Python libraries like Flask and Django to do your deployment instead of having to worry about using a Julia project that is still relatively young compared to those two options.

A disadvantage to this is that we will get an Array of DataFrame rows as a return.

In the unstable version of Lathe we will need to have the at parameter set (the float set to 75 percent), but in any stable version this is not the case.