Check 26 compatible issue with train model on latest release with using picamera

General
1

hailo001@hailo001:~/picamera2/examples/hailo $ python detect.py --model overlap.hef --labels labels.txt
[HailoRT] [error] CHECK failed - HEF file length does not match
[HailoRT] [error] CHECK_SUCCESS failed with status=HAILO_INVALID_HEF(26)
[HailoRT] [error] Failed parsing HEF file
[HailoRT] [error] Failed creating HEF
[HailoRT] [error] CHECK_SUCCESS failed with status=HAILO_INVALID_HEF(26)

hi after update to hailort 4.18 with hailo docker suite 2024-07.1 im still getting this compatible issues with a custom model train and one is there any update on the yolo model for yolov8n-obb

2

Hey @SAN,

This issue is related to a mismatch in the DFC version during compilation. We made a change after DFC 3.27. If you’re using HailoRT 4.18 with DFC version 3.27, you’ll need to recompile it using DFC 3.28 to resolve the problem.

3

ermm if that the case here a situation, i retrain a yolov8n-obb and convert it using hailo suite Hailo AI Software Suite – Docker 2024-07.1 and the on my raspberry pi i update hailo api using pip install hailo-all currently its running on hailort 4.18

image

does that considered as mismatch

4

Yes, that is the correct way to proceed. The 2024.07 Docker version includes DFC 3.28, so there will be no mismatches.

5

in that case why im getting the compatible issue check 26 doesnt make sense

6

Can you run the following command in the terminal:

Hailortcli parse-hef <yourmodel.hef>

and share the output with me so we can check what the issue might be?

8

There you go

image
image1394×323 97.5 KB

9

The .hef file looks fine. Now, could you run the hailo -h command inside the DFC environment or the Docker container where DFC is installed? This will display all the version details."

10

image
image1402×218 91.7 KB

yep here you go