How to Use Model Cascade in Mech-DLK?

In deep learning, if you need to perform Instance Segmentation to segment objects first and then use the Classification module to distinguish the fronts and backs of these objects, you can employ a model cascade.

How does such a model cascade work? Is there any documentation available?


  1. After training the instance segmentation model, click on “Validate”.
  2. Add the Classification module and import data from the previous level.
  3. After training the Classification model, click on “Export”.

For more information, please refer to the documentation center: Cascade Modules.