Sameer Sheorey

Hi @cgsaxner thanks for reporting this. Can you try with an empty `runs` directory? Can you provide the minimal contents of the `runs` directory that slows down tensorboard?

On Windows, `tensorboard` does not use fast data loading by default (`--load_fast=true`) https://github.com/tensorflow/tensorboard/issues/4784 I suspect this may be amplified by the Open3D plugin that needs to check each run to...

I believe this is technically possible with `pybind11`, but we don't explicitly support adding functionality to pybind classes in Python. This looks like a good feature to have. @dd-ff did...

Great, thanks Timo. I don’t have permission to update the PR, so will wait for you. Regards, Sameer From: Timo Röhling ***@***.***> Date: Tuesday, July 12, 2022 at 11:46 PM...

Hi @Nireil thanks for reporting this error. If you are getting a seg fault, please convert to a bug report, i.e. provide enough system / environment information for us to...

Hi @Yinxq the mapping between class label and class names for SemanticKITTI is here: https://github.com/isl-org/Open3D-ML/blob/5baf7223e9126e26feffc6f08d4b9ae728367dd1/ml3d/datasets/semantickitti.py#L112 You can modify this as appropriate for Rellis-3D.

Hi @KiviWong for remote Linux servers, you have these options: - Use CPU rendering with the environment variable `OPEN3D_CPU_RENDERING=True` . Slower than with - Use [VirtualGL](https://virtualgl.org/) `vglrun open3d_app_or_python_script`. Provides full...

Hi @roehling can you update the PR with master and look at the CI errors (pushing an update will rerun CI)? I will be happy to fix the CI errors...

Hi @roehling can you take a look at the Windows CI errors?

Closing in favor of adopted PR #5892.