search

drprojects / superpoint_transformer

Official PyTorch implementation of Superpoint Transformer introduced in [ICCV'23] "Efficient 3D Semantic Segmentation with Superpoint Transformer" and SuperCluster introduced in [3DV'24 Oral] "Scalable 3D Panoptic Segmentation As Superpoint Graph Clustering"
MIT License
567 stars 74 forks source link

Error when installing pgeof. #111

Closed Wind010321 closed 4 months ago

Wind010321 commented 4 months ago

Hello, when I build the environment following the install.sh,I encounter an error as follows:

截屏2024-05-25 13 00 54

I have tried many times but not find the solutions. My python version is 3.8 and the torch's version is 2.2.0, cuda's version is 11.8. Can you give me some solutions? Thanks!

drprojects commented 4 months ago

This may be related to #102. Can you please try the tentative fixes proposed in the thread and let me know if that solves your problem ?

Wind010321 commented 4 months ago

Sorry, the command "conda install -c conda-forge libstdcxx-ng" does not solve my problem.And the error proposed is the same as the original one.

Wind010321 commented 4 months ago

I just want to use the partition level L1 and L2 superpoints as the input of my network. But at the beginning of the installation part, I get the error. Is there any other solution that you can provide?

Wind010321 commented 4 months ago

I noticed that the supported python version of pgeof >=3.9. Should I change the python's version to 3.9 and install from scratch?

drprojects commented 4 months ago

Ok can you please try installing only pgeof in a fresh environment ? Refer to https://pypi.org/project/pgeof for that. If the error persists, this is a pgeof-specific error. In which case please open an issue on: https://github.com/drprojects/point_geometric_features

Wind010321 commented 4 months ago

It is happy that I successfully install the pgeof using command "python -m pip install pgeof".

Yesterday I have used this command line but failed to install it. I will figure it out in the last hours to give you some possible reasons.

Wind010321 commented 4 months ago

Hi, I find out that the clean conda environment is the key. Use the command "conda clean" might help. I have solve my problem and I will close the issue. Thanks for your patience!

Avril-Dragon commented 2 months ago

I also encountered the same difficulty and could not solve

Wind010321 commented 2 months ago

I have created this conda environment for more that 4 times. Just run the ./install.sh and you will make it!

I have created this conda environment for more that 4 times. Just run the ./install.sh and you will make it!