Skip to content

Latest commit

 

History

History
30 lines (18 loc) · 957 Bytes

CASCADE.rst

File metadata and controls

30 lines (18 loc) · 957 Bytes

Cascade

This document provides instructions and basic usage examples for inferring cascade networks.

General

  • The Cascade command enables the generating of sequence of models by cascading them one after the other to generate complex deep learning pipelines. For example, 3D object detection in bird's-eye-view pipeline such as PETRv2.
  • Currently the cascade eval API supports PETRv2 only, see petrv2_repvggB0.yaml for further configurations.
  • The user will need existing HARs/HEFs: both petrv2_repvggB0_backbone_pp_800x320 & petrv2_repvggB0_transformer_pp_800x320.

Evaluation

To evaluate a cascade model on different targets, use the cascade flag:

hailomz cascade eval petrv2
hailomz cascade eval petrv2 --override target=emulator
hailomz cascade eval petrv2 --override target=hardware

To explore other options use:

hailomz cascade eval --help