Systems and Methods for Producing an Architecture of a Pyramid Layer

A computing system comprising:

• one or more processors;
• a machine-learned model configured to receive an input, and in response to receiving the input, output an output, the machine-learned model comprising a backbone model and a plurality of pyramid layers stacked in a sequence one after the other, wherein the backbone model is configured to receive the input and to generate an initial plurality of output feature representations, wherein each pyramid layer of the plurality of pyramid layers is configured to receive a respective plurality of input feature representations that have a plurality of different resolutions from the backbone model or a previous sequential pyramid layer and to generate a respective plurality of output feature representations that have the same plurality of different resolutions;
• one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations, the operations comprising:
• inputting the input into the machine-learned model;
• and
• receiving, as an output of the machine-learned model, the output.