search

w111liang222 / lidar-slam-detection

LSD (LiDAR SLAM & Detection) is an open source perception architecture for autonomous vehicle/robotic
Apache License 2.0
457 stars 105 forks source link

Cannot see the Config UI #8

Open CharlieV5 opened 1 year ago

CharlieV5 commented 1 year ago

Hi, I'm trying to test your program following the instructions here : https://github.com/w111liang222/lidar-slam-detection/blob/main/docs/slam.md When I launch the LSD and open localhost:config , I cannot change the setting about mapping & localization. I just see the waiting page, as the picture below 2023-07-22 17-27-22 的屏幕截图 2023-07-22 17-28-04 的屏幕截图

w111liang222 commented 1 year ago

Hi, What's the gpu card you use? The program was blocked when import the slam module.

You can try the "15liangwang/auto-ipu" docker image which is the non-cuda version

CharlieV5 commented 1 year ago

The GPU is RTX2060. The nVidia drivers and cuda have been installed. Maybe I will try non-cuda version, later.

w111liang222 commented 1 year ago

The SM80+ GPU is required by libspconv and the RTX2060 is SM76, you can check it from https://developer.nvidia.com/cuda-gpus. So, we recommend to use the non-cuda version to test the LSD.

CharlieV5 commented 1 year ago

The SM80+ GPU is required by libspconv and the RTX2060 is SM76, you can check it from https://developer.nvidia.com/cuda-gpus. So, we recommend to use the non-cuda version to test the LSD.

Thank you! I'll try later.

CharlieV5 commented 1 year ago

Hi, @w111liang222 trying to use non-cuda docker, but now failed to start the docker container: 2023-08-07 20-54-18 的屏幕截图

w111liang222 commented 1 year ago

Sorry, I didn't encounter this problem, did you remove the "--gpus all" when you start the container

CharlieV5 commented 1 year ago

Sorry, I didn't encounter this problem, did you remove the "--gpus all" when you start the container

With or without "--gpus all" are both tried. 2023-08-08 17-30-46 的屏幕截图

w111liang222 commented 1 year ago

It is like an issue of nvidia docker, this link maybe can help you https://github.com/NVIDIA/nvidia-container-toolkit/issues/154

RuiYangQuan commented 1 year ago

The SM80+ GPU is required by libspconv and the RTX2060 is SM76, you can check it from https://developer.nvidia.com/cuda-gpus. So, we recommend to use the non-cuda version to test the LSD.

Hi, I'm experiencing the same problem, when I open localhost:config GUI keeps loading without displaying, both cuda version and non-cuda version I've tried.

XinyuMeng commented 1 year ago

Hello, would you like to ask if your problem has been solved? I would like to share your experience. Because I encountered a similar problem to yours, my GPU model is RTX3070, and the error interface is the same as yours. Looking forward to your reply

w111liang222 commented 1 year ago

Hello, this issue maybe caused by the host software environment, the details of the host that we test LSD are:

You can check it if the version is mismatch

XinyuMeng commented 1 year ago

Hello, this issue maybe caused by the host software environment, the details of the host that we test LSD are:

  • Ubuntu18.04
  • docker-compose/bionic,now 1.17.1-2 all
  • nvidia-docker2/bionic 2.13.0-1 all
  • GPU: RTX3090

You can check it if the version is mismatch

Hello, below is my environment information. It seems that only the GPU version is inconsistent with yours. I would like to ask you, how should I solve this problem next? Are there any examples of successful running of your code on the GPU model RTX3070? Thanks again

w111liang222 commented 1 year ago

Hello, this issue maybe caused by the host software environment, the details of the host that we test LSD are:

  • Ubuntu18.04
  • docker-compose/bionic,now 1.17.1-2 all
  • nvidia-docker2/bionic 2.13.0-1 all
  • GPU: RTX3090

You can check it if the version is mismatch

Hello, below is my environment information. It seems that only the GPU version is inconsistent with yours. I would like to ask you, how should I solve this problem next? Are there any examples of successful running of your code on the GPU model RTX3070? Thanks again

  • nvidia-docker2 2.13.0-1 all nvidia-docker CLI wrapper
  • docker-compose 1.17.1-2 all
  • docker-compose-plugin 2.18.1-1~ubuntu.18.04~bionic amd64
  • Ubuntu 18.04.6 LTS

I don't have RTX3070, so I only test the lsd with RTK3090. But the RTX3090 is similar to 3070, lsd should be able to run on both. Did you try the non-cuda docker image (15liangwang/auto-ipu) to check if the problem is caused by the GPU

XinyuMeng commented 1 year ago

Hello, this issue maybe caused by the host software environment, the details of the host that we test LSD are:

  • Ubuntu18.04
  • docker-compose/bionic,now 1.17.1-2 all
  • nvidia-docker2/bionic 2.13.0-1 all
  • GPU: RTX3090

You can check it if the version is mismatch

Hello, below is my environment information. It seems that only the GPU version is inconsistent with yours. I would like to ask you, how should I solve this problem next? Are there any examples of successful running of your code on the GPU model RTX3070? Thanks again

  • nvidia-docker2 2.13.0-1 all nvidia-docker CLI wrapper
  • docker-compose 1.17.1-2 all
  • docker-compose-plugin 2.18.1-1~ubuntu.18.04~bionic amd64
  • Ubuntu 18.04.6 LTS

I don't have RTX3070, so I only test the lsd with RTK3090. But the RTX3090 is similar to 3070, lsd should be able to run on both. Did you try the non-cuda docker image (15liangwang/auto-ipu) to check if the problem is caused by the GPU

Yes, I have tested the non-cuda docker image (15liangwang/auto-ipu) version and it can be displayed normally.

w111liang222 commented 12 months ago

@XinyuMeng Hi, did you solve this issue(no responce when use the cuda version image)? If yes, can you tell me how to solve it because others also encounter this issue.