search

renezurbruegg / icg_net

Implementation of the Paper: ICG-Net: A unified approach for instance centric grasping
MIT License
7 stars 2 forks source link

Conflict between cuda version and gcc version #3

Open Jinbo-Zuo opened 2 months ago

Jinbo-Zuo commented 2 months ago

In "conda_cu113.yml", you installed cuda11.3 and gcc11.2, but in fact, they are conflicting. When I install "pyg_lib torch_scatter torch_sparse torch_cluster torch_spline_conv", the error occur: RuntimeError: The current installed version of /home/username/anaconda3/envs/icg_cuda113/bin/x86_64-conda-linux-gnu-c++ (11.2.0) is greater than the maximum required version by CUDA 11.3 (10.0.0). Please make sure to use an adequate version of /home/username/anaconda3/envs/icg_cuda113/bin/x86_64-conda-linux-gnu-c++ (>=5.0.0, <=10.0.0). Here is the version correspondence table of cuda and gcc er

Jinbo-Zuo commented 2 months ago

In fact, conda_cu121.yml also cannot successfully create the desired conda environment. This is because cudatoolkit=12.1 is not yet supported in conda. ww

renezurbruegg commented 2 months ago

I just updated the conda_cu121.yml with my working local version. Could you check if the installation for this environment works for you?