Cv2.error: no CUDA-capable device is detected in function ‘ManagedPtr’

jens_g · August 28, 2023, 1:14pm

Hi!

My environment:

Jetson Nano
Jetpack 4.6.4
L4T: 32.7.4
CUDA 10.2.300
cuDNN 8.2.1.32
OpenCV 4.5.1 compiled with CUDA
Python 3.7.5

I’m trying to use OpenCV 4.5.1 with CUDA on Python. But I get this error:

cv2.error: OpenCV(4.5.1) /home/jetson/opencv/modules/dnn/src/cuda4dnn/csl/memory.hpp:54: error: (-217:Gpu API call) no CUDA-capable device is detected in function ‘ManagedPtr’

How can I solve it and make it work?
Thanks a lot.

Addition: same problem with OpenCV 4.4.0:

cv2.error: OpenCV(4.4.0) /home/jetson/opencv/modules/dnn/src/cuda4dnn/csl/memory.hpp:54: error: (-217:Gpu API call) no CUDA-capable device is detected in function ‘ManagedPtr’

Any idea?

jens_g · August 30, 2023, 6:24am

No reaction from support?

AastaLLL · August 30, 2023, 6:27am

Hi,

Sorry for the late update.

Please install it from the source to get the CUDA support.
Below is a script for your reference:

github.com

AastaNV/JEP/blob/master/script/install_opencv4.6.0_Jetson.sh

#!/bin/bash
#
# Copyright (c) 2022, NVIDIA CORPORATION.  All rights reserved.
#
# NVIDIA Corporation and its licensors retain all intellectual property
# and proprietary rights in and to this software, related documentation
# and any modifications thereto.  Any use, reproduction, disclosure or
# distribution of this software and related documentation without an express
# license agreement from NVIDIA Corporation is strictly prohibited.
#

version="4.6.0"
folder="workspace"

set -e

for (( ; ; ))
do
    echo "Do you want to remove the default OpenCV (yes/no)?"
    read rm_old

This file has been truncated. show original

Thanks.

jens_g · August 30, 2023, 2:24pm

Hi, I installed opencv 4.6.0 with the script you provided but something seems to be wrong. See below output. He complains about missing libopencv_hdf.so.4.5. How to fix it?

jetson@nano:~$ python
Python 3.7.5 (default, Dec 9 2021, 17:04:37)
[GCC 8.4.0] on linux
Type “help”, “copyright”, “credits” or “license” for more information.

import cv2
Traceback (most recent call last):
File “”, line 1, in
File “/usr/lib/python3/dist-packages/cv2/init.py”, line 129, in
bootstrap()
File “/usr/lib/python3/dist-packages/cv2/init.py”, line 112, in bootstrap
import cv2
ImportError: libopencv_hdf.so.4.5: cannot open shared object file: No such file or directory

AastaLLL · August 31, 2023, 8:43am

Hi,

Do you use a custom Python version? (3.7?)
If yes, please check if all the dependencies are updated for v3.7 as well.

Thanks.

jens_g · August 31, 2023, 8:59pm

Hi, I switched to Python3.8 and have compiled OpenCV 4.6.0 again based on the script you provided. Compilation and installation of OpenCV was successful, see test below:

jetson@nano:~$ python
Python 3.8.0 (default, Dec 9 2021, 17:53:27)
[GCC 8.4.0] on linux
Type “help”, “copyright”, “credits” or “license” for more information.

import cv2
cv2.version
‘4.6.0’
count = cv2.cuda.getCudaEnabledDeviceCount()
print(count)
1

But the original error from the beginning remains unchanged:

08/31/23 23:09:50 zmesdetect_m1[9928] ERR detect_sequence.py:713 [Error running model: OpenCV(4.6.0) /home/jetson/workspace/opencv-4.6.0/modules/dnn/src/cuda4dnn/csl/memory.hpp:54: error: (-217:Gpu API call) no CUDA-capable device is detected in function ‘ManagedPtr’]

08/31/23 23:09:50 zmesdetect_m1[9928] DBG2 detect_sequence.py:714 [Traceback (most recent call last):
File “/usr/local/lib/python3.8/dist-packages/pyzm/ml/detect_sequence.py”, line 710, in detect_stream
_b,_l,_c,_m = m.detect(image=frame)
File “/usr/local/lib/python3.8/dist-packages/pyzm/ml/object.py”, line 71, in detect
b,l,c,_model_names = self.model.detect(image)
File “/usr/local/lib/python3.8/dist-packages/pyzm/ml/yolo.py”, line 189, in detect
outs = self.net.forward(ln)
cv2.error: OpenCV(4.6.0) /home/jetson/workspace/opencv-4.6.0/modules/dnn/src/cuda4dnn/csl/memory.hpp:54: error: (-217:Gpu API call) no CUDA-capable device is detected in function ‘ManagedPtr’]

08/31/23 23:09:50 zmesdetect_m1[9928] DBG2 detect_sequence.py:770 [We did not find any object matches in frame: alarm]

08/31/23 23:09:50 zmesdetect_m1[9928] DBG2 yolo.py:82 [pyzm_uid33_gpu_lock portalock already released]

08/31/23 23:09:50 zmesdetect_m1[9928] DBG1 detect_sequence.py:830 [perf: TOTAL detection sequence (with image loads) took: 477.96 ms to process 550]

08/31/23 23:09:50 zmesdetect_m1[9928] DBG1 zm_detect.py:556 [Closing logs]

What else can I do to solve it? I’m totally desperate and happy to provide additional information, in case that would help to fix it. Please let me know.

AastaLLL · September 1, 2023, 7:35am

Hi,

For Nano, it’s more commended to use the default Python 3.6.

Is OpenCV 4.5.0 an option for you?
If yes, please try the below script which is verified under the JetPack 4 environment:

github.com

AastaNV/JEP/blob/master/script/install_opencv4.5.0_Jetson.sh

#!/bin/bash
#
# Copyright (c) 2020, NVIDIA CORPORATION.  All rights reserved.
#
# NVIDIA Corporation and its licensors retain all intellectual property
# and proprietary rights in and to this software, related documentation
# and any modifications thereto.  Any use, reproduction, disclosure or
# distribution of this software and related documentation without an express
# license agreement from NVIDIA Corporation is strictly prohibited.
#

version="4.5.0"
folder="workspace"

echo "** Remove other OpenCV first"
sudo sudo apt-get purge *libopencv*


echo "** Install requirement"
sudo apt-get update

This file has been truncated. show original

Thanks.

jens_g · September 1, 2023, 12:53pm

Hi, OpenCV 4.5.0 would also be OK but I need Python version beyond 3.6. Any idea how to make it work?

AastaLLL · September 4, 2023, 7:29am

Hi,

You will need to install the new Python version manually.
And update all the dependencies to the corresponding Python version.

We didn’t try it before but it should work since some users have successfully built it.
Thanks.

system · September 26, 2023, 8:41am

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.

Topic		Replies	Views
OpenCV error: no CUDA-capable device is detected in function 'ManagedPtr' Jetson Nano opencv , cuda	2	1474	October 18, 2021
Cannot import OpenCV4.6.0 from Python3.8 Jetson Nano opencv , cuda	9	477	May 7, 2024
OpenCV,CUDA,Python? Jetson Nano opencv	4	2043	October 14, 2021
Opencv incompatible Jetson Nano opencv	6	434	October 18, 2021
Opencv 4.6.0 with cuda support were successfully built for both python2.7 and python3.8, but cuda unavailable Jetson Xavier NX opencv , cuda , pytorch , cudnn	4	2773	November 2, 2023
Unable to install opencv with CUDA in Jetson Nano Jetson Nano	30	13772	October 18, 2021
OpenCV with CUDA error - device unavailable. Jetson Orin Nano Jetson Orin Nano opencv , cuda , cudnn	7	306	May 21, 2024
Error trying to use OpenCV with CUDA support on Docker: CUDA driver version is insufficient for CUDA runtime version Jetson Nano opencv , cuda , docker	4	2897	June 27, 2022
How to manually install CUDA and all necessary packages on my Jetson Nano without SDK Manager? Jetson Nano cudnn	7	8643	March 18, 2024
How to install opencv-python for python3.6 Jetson Nano opencv	26	19382	October 14, 2021

Cv2.error: no CUDA-capable device is detected in function ‘ManagedPtr’

Related topics