search

JonasSchult / Mask3D

Mask3D predicts accurate 3D semantic instances achieving state-of-the-art on ScanNet, ScanNet200, S3DIS and STPLS3D.
MIT License
525 stars 103 forks source link

How do I run this network with my own data? #146

Open mostafa501 opened 9 months ago

mostafa501 commented 9 months ago

Congrats on the release and thank you for open sourcing your code! I would love to know if you have any plans to open-source the inference code used on the demo you provide. Currently, it seems that only the train/test code is available, and would love to use a python interface to use this method with some different point clouds. Any help on this would be very much appreciated. Thank you for the contribution !!

nekoshadow1 commented 9 months ago

I think you may refer to a quick setup here: https://github.com/cvg/Mask3D, where you can feed a custom .ply file and get the visualizations. However, only Scannet200 checkpoints are supported for now.

mostafa501 commented 9 months ago

@nekoshadow1, thank you for the reply, please i have more questions: 1-can we use the above guide to test own collected data using CPU not GPU or not ? 2-can we use this guide with other pretrained models like scannet or s3dis or not ? Thank you so much.

nekoshadow1 commented 9 months ago

@nekoshadow1, thank you for the reply, please i have more questions: 1-can we use the above guide to test own collected data using CPU not GPU or not ? 2-can we use this guide with other pretrained models like scannet or s3dis or not ? Thank you so much.

1- The code relies on cuda 11.3 so it runs on GPU only. I have tried to install it in a CPU environment but failed. 2- For now the guide only works with ScanNet200 pretrained models. I think the author of the guide is working on ScanNet models but there is no ETA.