I am currently working on an academic project using the Hailo-8L module with a Raspberry Pi 5. I’m using the hailo-rpi5-examples repository to run real-time YOLO inference with camera input.
However, I encountered a segmentation fault during execution due to a missing shared object file:
libyolo_hailortpp_postprocess.so
The detection script looks for this file at:
/usr/local/hailo/resources/so/libyolo_hailortpp_postprocess.so
Unfortunately, this file is not included in the latest Hailo Model Zoo or the RPi examples. It also seems that the source code to build this .so is no longer publicly available.
Could you kindly provide one of the following:
- A working copy of libyolo_hailortpp_postprocess.so (compatible with Hailo-8L)
- Or access to the source files (e.g., CMake + .cpp) to compile it myself
This is currently blocking me from completing the integration of the video stream and post-processed inference.
For reference, I’ve already installed:
- hailort_4.21.0_arm64.deb
- hailort-pcie-driver_4.21.0_all.deb
- python3-hailort_4.21.0-1_arm64.deb
Thanks a lot for your support!
Raspberry Pi 5 + Hailo-8L user
(Individual developer / academic usage)
This is our code
import sys
sys.path.insert(0, '/home/ihatakimi/yolo_clean_env/lib/python3.11/site-packages/hailort-4.20.0-py3.11-linux-aarch64.egg')
from hailo_platform.pyhailort import pyhailort
import numpy as np
import time
import cv2
# Hailo alias'ları
HEF = pyhailort.HEF
Device = pyhailort.Device
HailoStreamInterface = pyhailort.HailoStreamInterface
def preprocess_image(image_path, input_shape):
"""Görüntüyü modele uygun hale getir"""
img = cv2.imread(image_path)
img = cv2.resize(img, (input_shape[2], input_shape[1]))
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
return img
def main():
print("📸 Hailo Algılama ve FPS Testi Başlıyor...")
try:
# HEF ve görüntü yolu
hef_path = "yolov8s.hef"
image_path = "test.jpg"
# HEF yükle
hef = HEF(hef_path)
print("✅ HEF yüklendi")
# Cihazla yapılandır
with Device() as device:
network_group = device.configure(hef, interface=HailoStreamInterface.PCIe)
print("✅ Cihaz yapılandırıldı")
# Stream bilgileri
input_stream = network_group.get_input_streams()[0]
input_shape = input_stream.shape
print(f"📐 Girdi şekli: {input_shape}")
# Görüntüyü hazırla
img = preprocess_image(image_path, input_shape)
img_np = np.expand_dims(img, axis=0).astype(np.uint8) # (1, H, W, C)
# Isınma
for _ in range(5):
network_group.run({input_stream: img_np})
print("🔥 Isınma tamamlandı")
# FPS Testi
num_frames = 50
start = time.time()
for _ in range(num_frames):
network_group.run({input_stream: img_np})
end = time.time()
fps = num_frames / (end - start)
print(f"\n🎯 Ortalama FPS: {fps:.2f}")
except Exception as e:
print(f"❌ Hata oluştu: {e}")
if _name_ == "_main_":
main()
This is our error
Traceback (most recent call last):
File "/home/ihatakimi/yolo_clean_env/fps_test.py", line 1, in <module>
from hailort import HEF, Device, VDevice, HailoStreamInterface
ModuleNotFoundError: No module named 'hailort'
Activated the correct Python virtual environment (
We are kindly asking: