Skip to content

Latest commit

 

History

History
77 lines (45 loc) · 2.58 KB

README.rst

File metadata and controls

77 lines (45 loc) · 2.58 KB
Raw

x86 Accelerated Century Pipeline

Overview

century.sh demonstrates detection on one video file source over multiple Hailo-8 devices, either using the Century platform, or other multi device configurations (i.e., multiple M.2 modules connected directly to the same host). While this application defaults to 4 devices, any number of Hailo-8 devices are supported.
This pipeline runs the detection network Yolov5.

In this pipeline the decoding phase is accelerated by VA-API.

Options

./century.sh [--input FILL-ME]
  • --input is an optional flag, a path to the video displayed (default is detection.mp4).
  • --show-fps is an optional flag that enables printing FPS to the console.
  • --print-gst-launch is a flag that prints the ready gst-launch command without running it
  • --device-count is an optional flag that sets the number of devices to use (default 4)

Configuration

The app post process parameters can be configured by a json file located in $TAPPAS_WORKSPACE/apps/h8/gstreamer/general/century/resources/configs/yolov5.json

Supported Networks

Run

cd $TAPPAS_WORKSPACE/apps/h8/gstreamer/general/century
./century.sh

The output should look like:

How the application works

This app is based on our multi device pipeline template.

How to use Retraining to replace models

Note

It is recommended to first read the Retraining TAPPAS Models page.

You can use Retraining Dockers (available on Hailo Model Zoo), to replace the following models with ones that are trained on your own dataset:

  • yolov5m

    • For optimum compatibility and performance with TAPPAS, use for compilation the corresponding YAML file from above.
    • TAPPAS changes to replace model:
      • Update HEF_PATH on the .sh file
      • Update resources/configs/yolov5.json with your new post-processing parameters (NMS)