YOLOv5

YOLOv5 on reComputer R Series (CM5 + Hailo-10H)

YOLOv5s and YOLOv5m perform COCO 80-class object detection on Hailo-10H through HailoRT. NMS and sigmoid run on-chip (Hailo HPP), so the host only parses the post-NMS tensor.

This page targets reComputer R Series (CM5 + Hailo-10H) with a PCIe Hailo-10H accelerator.

Hardware and Host Setup

Install the Hailo-10H packages on the host and confirm that the driver and device are ready before starting the container:

sudo apt update
sudo apt install hailo-h10-all -y
sudo reboot

# After reboot
hailortcli --version
hailortcli fw-control identify
ls -l /dev/hailo0

The expected HailoRT version for these images is 5.1.1.

Run With Demo Video

YOLOv5s:

sudo docker run --rm \
  --name cm5-hailo10h-yolov5s \
  --privileged \
  --net=host \
  -e PYTHONUNBUFFERED=1 \
  --device /dev/hailo0:/dev/hailo0 \
  -v /usr/lib/libhailort.so.5.1.1:/usr/lib/libhailort.so.5.1.1:ro \
  -v /usr/lib/libhailort.so:/usr/lib/libhailort.so:ro \
  ghcr.io/seeed-projects/recomputer-hailo10h-cv/yolov5s:latest \
  python web_detection.py --model_path model/yolov5s.hef --video_path video/test.mp4

YOLOv5m:

sudo docker run --rm \
  --name cm5-hailo10h-yolov5m \
  --privileged \
  --net=host \
  -e PYTHONUNBUFFERED=1 \
  --device /dev/hailo0:/dev/hailo0 \
  -v /usr/lib/libhailort.so.5.1.1:/usr/lib/libhailort.so.5.1.1:ro \
  -v /usr/lib/libhailort.so:/usr/lib/libhailort.so:ro \
  ghcr.io/seeed-projects/recomputer-hailo10h-cv/yolov5m:latest \
  python web_detection.py --model_path model/yolov5m.hef --video_path video/test.mp4

Open http://<Board_IP>:8000 to view the web preview.

USB Camera Mode

sudo docker run --rm \
  --name cm5-hailo10h-yolov5s \
  --privileged \
  --net=host \
  -e PYTHONUNBUFFERED=1 \
  --device /dev/hailo0:/dev/hailo0 \
  --device /dev/video0:/dev/video0 \
  -v /usr/lib/libhailort.so.5.1.1:/usr/lib/libhailort.so.5.1.1:ro \
  -v /usr/lib/libhailort.so:/usr/lib/libhailort.so:ro \
  ghcr.io/seeed-projects/recomputer-hailo10h-cv/yolov5s:latest \
  python web_detection.py --model_path model/yolov5s.hef --camera_id 0

Swap yolov5s for yolov5m to use the medium variant.

HEF Files

YOLOv5s is the small, YOLOv5m the medium variant; both are Hailo Model Zoo v5.4.0 Hailo-10H builds.

Source: https://hailo-model-zoo.s3.eu-west-2.amazonaws.com/ModelZoo/Compiled/v5.4.0/hailo10h/yolov5s.hef (yolov5m.hef in the same directory).

REST API

POST http://<Board_IP>:8000/api/models/yolov5s/predict

curl -X POST "http://<Board_IP>:8000/api/models/yolov5s/predict" \
  -F "file=@test.jpg"

Implementation Notes

• The HEF runs NMS and sigmoid on-chip (zoo base/yolo.yaml: nms: true, hpp: true), so nms_thresh is accepted for API parity but ignored.
• Post-NMS rows are [ymin, xmin, ymax, xmax, score], normalized to [0,1] of the 640x640 letterboxed input; the app scales to pixels and un-letterboxes.
• The HEF bakes normalize_in_net (mean 0 / std 255, a /255 scale) and padding_color=114: the app letterboxes with gray (114) and feeds raw uint8 RGB pixels — no manual normalization.
• Class IDs 0..79 index the standard COCO class list directly (the Model Zoo evaluation uses labels_offset=1).
• The Model Zoo documents the post-NMS shape as 80x5x80; the parser accepts the real layout at runtime and logs it once ([YOLOv5] raw output type=..., shape=...) for on-device verification.
• Defaults: confidence 0.25; the IOU slider has no effect because NMS is on-chip.

Development Notes

• Source modules: src/hailo10h_yolov5s/, src/hailo10h_yolov5m/
• Dockerfiles: docker/hailo10h/yolov5s.dockerfile, docker/hailo10h/yolov5m.dockerfile
• Containers: ghcr.io/seeed-projects/recomputer-hailo10h-cv/yolov5s:latest, .../yolov5m:latest
• Both modules use the HailoRT 5.1.1 InferModel API and Hailo-10H-specific HEFs.
• Unit tests run without a device: python -m unittest discover -s tests -v.