exit code -11 while running darknet_ros.launch

[Looomo]

error lookes below:

[darknet_ros-1] process has died [pid 23440, exit code -11, cmd /home/shanyx/catkin_ws/devel/lib/darknet_ros/darknet_ros camera/rgb/image_raw:=/image_raw __name:=darknet_ros __log:=/home/shanyx/.ros/log/5c4f5838-3449-11ea-8e87-54e1ad5a8195/darknet_ros-1.log]. log file: /home/shanyx/.ros/log/5c4f5838-3449-11ea-8e87-54e1ad5a8195/darknet_ros-1*.log

i have checked the logfile , nothing useful . Can anyone please help me ?

[Looomo]

when i use gdb , ii got this :

[New Thread 0x7fff369fc700 (LWP 12457)]
Waiting for image.
[New Thread 0x7fff30dfe700 (LWP 12499)]

Thread 1 "darknet_ros" received signal SIGSEGV, Segmentation fault.
0x00007ffff7263e53 in ros::CallbackQueue::callAvailable(ros::WallDuration) ()
   from /opt/ros/lunar/lib/libroscpp.so

can anyone please help?

[Scarf195]

Hi, same problem, did you fix that?

[Looomo]

Hi, same problem, did you fix that?

no,and i found something strange . When I used Ubuntu 16 in VMwrea I can conmpile and run perfectly (CPU version) , but when i install Ubuntu on my laptop with the same ISO , it shows errors . I have changed the kernel to the version i used in VMware .

[Looomo]

Hi, same problem, did you fix that?

Maybe i fix that . I reinstalled ubuntu16 today , but i didnt install boost or opencv , just ROS kinetic , now this project worked with CPU . I havent test the GPU version . Hope this is helpful to you .

[Looomo]

Hi, same problem, did you fix that?

Also remember to change the path of config and weights files in YoloObjectDetector.cpp .

[Litao917]

Hey guys, I met the same issue. I'm on Ubuntu 18.04 with CUDA 10.2. Any clue about it? I currently would not like to reinstall my ROS as well as my Ubuntu. It is too costly for me. Many thanks!!!!

[darkhan-s]

Hey guys, I met the same issue. I'm on Ubuntu 18.04 with CUDA 10.2. Any clue about it? I currently would not like to reinstall my ROS as well as my Ubuntu. It is too costly for me. Many thanks!!!!

Hi, have you solved this problem?

[tomlankhorst]

I hope a more recent version of darknet_ros solves this issue. If not, feel free to reopen this issue.

[hanm2019]

I meet the same problem in melodic branch with yolov3.launch at ubuntu 18.04 ros melodic

I even try with the master branch, get the same error @tomlankhorst

[darkhan-s]

I think what I did back then to resolve this is that I retrained the model

[tomlankhorst]

Maybe try to download the pre-trained models to be sure.

[hanm2019]

Maybe try to download the pre-trained models to be sure.

I try with pre-trained models , but meet the same error

[dscroft]

Check the image format of the incoming message. I was getting the same error but traced it to the rgbgr_image function when I was passing in mono images. Switching to bgra8 caused the same exit code error but elsewhere.

Looks like only 3 channel image formats work but I'm still digging into it.

[hanm2019]

Check the image format of the incoming message. I was getting the same error but traced it to the rgbgr_image function when I was passing in mono images. Switching to bgra8 caused the same exit code error but elsewhere.

Looks like only 3 channel image formats work but I'm still digging into it.

I use ros_bridge to reform the image msg by:

cv_img_h = bridge.cv2_to_imgmsg(img_rot,"rgb8")

maybe the image has 3 channel

[TeingQiHao]

Loading weights from /home/bojan/darknet_ros_og/src/darknet_ros/darknet_ros/yolo_network_config/weights/yolov4-tiny-custom_2000.weights...Done! Waiting for image. [darknet_ros-1] process has died [pid 13861, exit code -11, cmd /home/bojan/darknet_ros_og/devel/lib/darknet_ros/darknet_ros camera/rgb/image_raw:=/camera/rgb/image_raw __name:=darknet_ros __log:=/home/bojan/.ros/log/869f03e4-4e65-11ec-b140-78d0042abcfe/darknet_ros-1.log]. log file: /home/bojan/.ros/log/869f03e4-4e65-11ec-b140-78d0042abcfe/darknet_ros-1*.log all processes on machine have died, roslaunch will exit shutting down processing monitor... ... shutting down processing monitor complete done

Hi everyone,

I am facing the same issue. I've trained a custom model on darknet yolov4-tiny. Followed the instructions on the readme.

1. Copied .weights file to darknet_ros/yolo_network_config/weights
2. Copied .cfg file to darknet_ros/yolo_network_config/cfg
3. Changed ros.yaml, camera reading topic to /zed/zed_node/left/image_rect_color (using zed_ros_wrapper)
4. Changed yolov2-tiny.yaml to point to config files and weights file per 1. and 2.
5. Changed detection classes names in yolov2-tiny.yaml to the 1 class the model was trained on ("- box")

I was able to run the default yolov2-tiny.weights and .cfg but face this error only after pointing to our .cfg and weights as mentioned.

Please help.

[MBouallegue]

This package helped me to use the yolov4 and worked with my custom dataset https://github.com/Tossy0423/darknet_ros/tree/020b7560f9c5901605503bb59450ed7654514e9c I guess in the old version of darknet (melodic for me) some elements in the darknet/src is not updated to suits the yolov4

[idochi]

I have the same problem and couldn't solve it yet. I'm using cv_camera for the image_raw. I'm new to Ros and need help.

Loading weights from /home/ez10/catkin_ws/src/darknet_ros/darknet_ros/yolo_network_config/weights/yolov3.weights...Done! Waiting for image. [darknet_ros-1] process has died [pid 9663, exit code -11, cmd /home/ez10/catkin_ws/devel/lib/darknet_ros/darknet_ros camera/rgb/image_raw:=dev/video0 __name:=darknet_ros __log:=/home/ez10/.ros/log/eed6e908-aff5-11ec-8912-4f219ac33992/darknet_ros-1.log]. log file: /home/ez10/.ros/log/eed6e908-aff5-11ec-8912-4f219ac33992/darknet_ros-1*.log all processes on machine have died, roslaunch will exit shutting down processing monitor... ... shutting down processing monitor complete done

[soroush-msd]

Hi, I'm also having the same problem. I'm running on Ubuntu 20.04 ARM64 with ROS Noetic. Any help is appreciated.

Loading weights from /home/parallels/catkin_ws/src/darknet_ros/darknet_ros/yolo_network_config/weights/yolov3.weights...Done! Waiting for image. [darknet_ros-1] process has died [pid 26142, exit code -11, cmd /home/parallels/catkin_ws/devel/lib/darknet_ros/darknet_ros camera/rgb/image_raw:=camera/rgb/image_raw __name:=darknet_ros __log:=/home/parallels/.ros/log/a2c43014-b08d-11ec-b516-814ad1a9efbb/darknet_ros-1.log]. log file: /home/parallels/.ros/log/a2c43014-b08d-11ec-b516-814ad1a9efbb/darknet_ros-1*.log all processes on machine have died, roslaunch will exit shutting down processing monitor... ... shutting down processing monitor complete done

[idochi]

The only thing, that worked for me was a new installation of Linux and ROS.

[AhmetEkiz]

Hi guys, I have the same problem, but it also says:

CUDA Error: no kernel image is available for execution on the device

Before that, I deleted all -gencode arch=compute_30,code=sm_30 like below on on CMakeLists.txt according to this solution.

Then I change the line as a -gencode arch=compute_75,code=sm_75 according to this solution. And It works.

Also, I found a similar solved issue.

System (please complete the following information):

OS: [Ubuntu 20.04]
ROS version: [Noetic]
GPU : [ GTX 1650]
Cuda: 12.1
Nvidia Driver Version: 530.30.02

-gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_50,code=[sm_50,compute_50] -gencode arch=compute_52,code=[sm_52,compute_52] -gencode arch=compute_61,code=sm_61 -gencode arch=compute_62,code=sm_62

[Vamsi-IITI]

Try the following method : Here