About the HailoRT error message on Raspberry pi5 platform

General
1

Hello,
I created a custom trained yolo v7 model and converted it to a *.hef file using DFC. I tried running it on RPI5 on Hailo 8L platform but got the following error message.
“”’

[HailoRT] [error] CHECK failed - Error opening file resource/v7best.hef
[HailoRT] [error] CHECK_SUCCESS failed with status=HAILO_OPEN_FILE_FAILURE(13)
[HailoRT] [error] Failed parsing HEF file
[HailoRT] [error] Failed creating HEF
[HailoRT] [error] CHECK_SUCCESS failed with status=HAILO_OPEN_FILE_FAILURE(13)
[HailoRT] [error] CHECK_SUCCESS failed with status=HAILO_OPEN_FILE_FAILURE(13)
[HailoRT] [error] CHECK_SUCCESS failed with status=HAILO_OPEN_FILE_FAILURE(13)
CHECK_EXPECTED failed with status=13
Showing FPS
Error: gst-stream-error-quark: Internal data stream error. (1), …/libs/gst/base/gstbasesrc.c(3132): gst_base_src_loop (): /GstPipeline:pipeline0/GstFileSrc:src_0:
streaming stopped, reason error (-5)
Shutting down… Hit Ctrl-C again to force quit.

“”"
Does this mean the *.hef file is corrupted?
Or is it not translated well and is a bad hef file? What should I look into to fix it? Thanks foe your help.

2

@sxk0988 can not say exactly but i think you forgot to add hw-arch hailo8l during compilation.

3

@saurabh hello, thanks for your answer.
I re-do the parsing, optimize and compile tasks and make sure adding the " hw-arch hailo8l " into the compile code as following.
“”"
chosen_hw_arch = ‘hailo8l’
runner = ClientRunner(hw_arch=chosen_hw_arch, har=har_path)
“”"
But I got the same error message on RPI5.
What could cause problem form the message “CHECK_EXPECTED failed with status=13” for this issue?
The command I used on RPI5 is as followings. Would it cause this issue?
“”"
python3 basic_pipelines/detection.py --hef-path resource/v7best.hef --input resources/teststream.mp4 --show-fps
“”"

4

This issue may caused by the onnx version and opset version. Re-install the onnx version to 1.12 and opset as 12 then the issue been fixed.

1 Like