Marnonel6/YOLOv7_ROS2

The numpy model cannot be found

Mrkofly opened this issue · 21 comments

When I run ros2 launch object_detection object_detection.launch.xml use_YOLOv7:=True

I successfully ran the node but it did not publish any topics
Hope to receive your guidance and suggestions

Hi,

Did you follow the extra steps and dependencies section?

You have to move the '.pt' weights file as outlined in the extra steps section.

Copy and paste your terminal error code here.

Thank you. I have already resolved this issue
But I am currently facing a problem where when I successfully run it, it will only proceed to the following step:
Uploading 截图 2024-04-16 17-49-05.png…

It won't run down, it won't post topics. Also, I hope you can provide a version of the dependent software for your environment configuration, which requirement.txt. Hope to receive your guidance and suggestions

Your image does not display.

Sorry, here are my image
1

This is the running process:
(ros2)mrfMrf: /ros2 humbles ros2 launch object detection object detection.launch.xmluse RGB:=True use YoLOv7:=True[launch]: All log files can be found below /home/mrf/.ros/log/2024-04-16-[INFO]17-45-48-151148-Mrf-12557[INFo] [launch]: Default logging verbosity is set to INFo[INFO][object detection-1]: process started with pid [12558][object detection-1] YoLoR2023-3-18 torch 2.2.2+cu121 CPUobject detection-1]object detection-1]Model Summary: 208 layers, 6013008 parameters, 0 gradients
When I use the "ros2 topic list" to view topics, there are no topics related to objection_detection

/door /person /stairs are all 3 object detection topics. You can 'echo' them and if it is detected there will be data published on those topics.

Can you please explain in detail how to solve this? I'm a bit confused. Thank you for your guidance

I would like to know about the Python environment and configuration you were running in at that time. Can you also provide some information? Thank you very much

My problem has been resolved. Thank you for your guidance

I don't have this environment anymore as it was a while back this project. If you know the requirements for the environment please make a Pull request with a requirements.txt file and I'll approve it.

Okay, when I succeed, I will make a Pull

If you need any help let me know.

Hi ,I still can't solve the problem of not displaying topics
I would like to use a Zed camera to load the topic of/Zed/zed_node/left/image-rect_color. Can you help me with this? Thank you for your guidance

Your photo did not upload

sorry
Uploading 1.jpg…

The photo is the same as the photo in the sixth message. I can't upload the photo anymore. I'm sorry

This was done with an Intel realsense camera not a ZED. So you'll probably have to make changes in the code.

I still can't post topics when I don't use userealsense
Can you help me subscribe to the topic of zed/zed_node/left/image rect_color in launch.xml