Question about the configuration.

[Gatsby23]

Hello, earlier I consulted about D2SLAM configuration issues. According to your advice, I tried to run it using Dockerfile. However, after trying it on two computers, I found that the problem still exists. The specific error is related to the LibTorch version, and after my testing, I found that the version（https://download.pytorch.org/libtorch/cu113/libtorch-shared-with-deps-1.10.1%2Bcu113.zip ）is more suitable. However, when I am running the D2SLAM algorithm with a monocular fisheye camera, the algorithm prompts that it cannot receive the message, and I suspect this is related to the ROS communication method. Therefore, I am looking for relevant debugging and configuration methods. If you have any suggestions or solutions, please feel free to share them. Additionally, I would greatly appreciate it if you could update the Dockerfile for our direct use.

[Gatsby23]

[minje-KIM]

Hello, Gatsby23. Did you solve this message receiving issue? I have totally same issue.

[xuhao1]

@Gatsby23 @minje-KIM We are using LIBTORCH_VERSION=1.13.0 It has been fixed in latest patch, thanks for your feedback For the message issue, make sure is compressed_images is set to zero, otherwise, D2SLAM will try to read from some~topic/compressed

is_compressed_images: 0