Yolov8n_seg instead of Yolov5n_seg on Raspberry Pi | How to do this?

General Discussion
1

Hello Hailo Community,

We are currently performing instance_segmentation with the yolov5n_seg model, which is the only one available for TAPPAS, according to this Repo hailo_model_zoo/docs/public_models/HAILO8L/HAILO8L_instance_segmentation.rst at master · hailo-ai/hailo_model_zoo · GitHub
Now with the new updates of the whole Hailo-apps, will there be an update for this segmentation to use yolo8n instead of yolo5n?

Or is there another way for us to use the yolo8n instead of yolo5n?

Currently when I try to run it with yolo8n_seg I receive this error, when trying it with the example from the basic pipeline:

(venv_hailo_rpi_examples) mi@raspberrypi:~/LoggingTool/Client $ python hailo/hailo-rpi5-examples/basic_pipelines/instance_segmentation.py --hef-path /home/mi/Downloads/yolov8n_seg.hef

Loading environment variables from /home/mi/LoggingTool/Client/hailo/hailo-rpi5-examples/.env…

:white_check_mark: All required environment variables loaded.

Auto-detected Hailo architecture: hailo8l

Traceback (most recent call last):

File “/home/mi/LoggingTool/Client/hailo/hailo-rpi5-examples/basic_pipelines/instance_segmentation.py”, line 136, in 

app = GStreamerInstanceSegmentationApp(app_callback, user_data)

      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

File “/home/mi/LoggingTool/Client/hailo/hailo-rpi5-examples/venv_hailo_rpi_examples/lib/python3.11/site-packages/hailo_apps/hailo_app_python/apps/instance_segmentation/instance_segmentation_pipeline.py”, line 60, in _init_

raise ValueError("HEF version not supported; please provide a compatible segmentation HEF or config file.")

ValueError: HEF version not supported; please provide a compatible segmentation HEF or config file.

Can someone help me, how I can achieve this?
Or will there be a solution with the new Hailo-Apps v25.10.0?

2

Hey @RRichard,

Welcome to the Hailo Community!

This is likely caused by one of two issues, but you’ll probably need to address both:

  1. You need to set up the postprocessing for it - make sure you’ve specified the output format in the config file so the NMS postprocess can work properly.

  2. This model from version 2.17 of the model zoo is designed for Hailo8 with HailoRT version 4.23, so double-check that you’re running the right HailoRT version!

Let me know how it goes.