Converting YOLOX_S model to HEF format

General Discussion
, ,
1

Hey,

I was able to use hailomz to convert my yolox_s.onnx file to yolox_s.hef, by following documentation using:

hailomz compile --ckpt yolox_s.onnx --calib-path ../yolov8n_files/car/ --yaml yolox_s.yaml --classes 80 --hw-arch hailo8

and it was successfull but when I try to run it, on my RPi 5, I get the following error

hailopi@raspberrypi:~/hailo_rnd $ hailortcli run yolox_s_self.hef
Running streaming inference (yolox_s_self.hef):
  Transform data: true
    Type:      auto
    Quantized: true
[HailoRT] [error] CHECK failed - Failed to extract_detections, reg yolox_s/conv55_111 buffer_size should be 25600, but is 6400
[HailoRT] [error] CHECK_SUCCESS failed with status=HAILO_INVALID_ARGUMENT(2)
[HailoRT] [error] CHECK_SUCCESS failed with status=HAILO_INVALID_ARGUMENT(2)
[HailoRT] [error] CHECK_SUCCESS failed with status=HAILO_INVALID_ARGUMENT(2)
Network yolox_s/yolox_s: 100% | 0 | FPS: 0.00 | ETA: 00:00:00
[HailoRT CLI] [error] Failed waiting for threads with status HAILO_INVALID_ARGUMENT(2)
[HailoRT CLI] [error] CHECK_SUCCESS failed with status=HAILO_INVALID_ARGUMENT(2)
[HailoRT CLI] [error] CHECK_SUCCESS failed with status=HAILO_INVALID_ARGUMENT(2)
[HailoRT CLI] [error] CHECK_SUCCESS failed with status=HAILO_INVALID_ARGUMENT(2)
[HailoRT CLI] [error] CHECK_SUCCESS failed with status=HAILO_INVALID_ARGUMENT(2) - Error failed running inference

But using similar method, I have been able to convert yolov8n.onnx file to yolov8n.hef format, and run it easily run command, so my installation is correct.

There has been another thread with same issue, which wasn’t answered: Converting YOLOX-s model to hef format - General - Hailo Community

2

Hi @Shashwat_Pandey1

Welcome to the Hailo community. Can you share the output of this command?
hailortcli parse-hef <your-model>.hef. Replace your-model string with the name of your model hef.

4

Here is the output-

hailopi@raspberrypi:~/hailo_rnd $ hailortcli parse-hef yolox_s_self.hef Architecture HEF was compiled for: HAILO8 Network group name: yolox_s, Multi Context - Number of contexts: 2     Network name: yolox_s/yolox_s         VStream infos:             Input  yolox_s/input_layer1 UINT8, NHWC(640x640x3)             Output yolox_s/yolox_nms_postprocess FLOAT32, HAILO NMS BY CLASS(number of classes: 80, maximum bounding boxes per class: 100, maximum frame size: 160320)             Operation:                 Op YOLOX                 Name: YOLOX-Post-Process                 Score threshold: 0.200                 IoU threshold: 0.65                 Classes: 80                 Cross classes: false                 NMS results order: BY_CLASS                 Max bboxes per class: 100                 Image height: 640                 Image width: 640