Model Pipeline #23

Background

The model pipeline describes our main entrypoint for creating clean patches from noise. Callers send in a noisy patch of fixed size, then get back a cleaned patch of the same size. Here, we load the models from #16 and pass inputs between the two models. This module will be called by the #4 for every region it cuts.

Description

To have a barebones workflow we should be able to give a patch and return a patch of the same size

• Load mock models from inference
• Connect model inputs and outputs to make a full resolution output
• Apply the elevation encoding and return the output