adkAurora

where is unlabeled class now?

> Hi! In DDP, `model.train_num_rays` and `model.max_train_num_rays` are defined **per device**, so you could either halving these values or simply train for fewer iterations :) Hi! I halved the parameters...

I met the same problem.

i met the same problem , i solved it by install taichi==0.7.13 `pip3 install taichi==0.7.13` `pip3 install taichi-tina` 

About the code you gave at [#282 ](https://github.com/DLR-RM/BlenderProc/issues/282), Where should this part of the code be modified? Is it in the [SegMapRendererUtility.py](https://github.com/DLR-RM/BlenderProc/blob/610aa3fceea748e2424db7285733508e45a17113/blenderproc/python/renderer/SegMapRendererUtility.py) file? Or is it simply a matter of...

> > Or is it simply a matter of incorporating this code into the main rendering code and utilizing it for rendering? > > `bproc.renderer.enable_segmentation_output(map_by=["cp_Roughness", "cp_Metallic", 'instance'],default_values={'roughness': None, 'metallic': None}...

Hey [cornerfarmer](https://github.com/cornerfarmer), I have a question. For models with roughness and metallic textures, can this method still be used? I tried it, but it seems to only return default values....