cdvae
cdvae copied to clipboard
Error when doing training: Segmentation fault
Hi professor, I followed the installation in README and use env.yaml to create the conda environment on GPU clusters. When i try command like python cdvae/run.py data=carbon
an error occurs:
[2022-12-15 21:18:19,363][hydra.utils][INFO] - Instantiating <cdvae.pl_data.datamodule.CrystDataModule> Segmentation fault (core dumped)
I have tried other data like mp_20 and tried to add CXX=g++
before command but still failed. Could you please help me about that? Thanks !
This is due PyTorch version. Reinstall the PyTorch correct version and it will work.
This is due PyTorch version. Reinstall the PyTorch correct version and it will work.
Hi professor, I encounter the same problem. Which PyTorch version should I install in the GPU cluster? .env setting or writing the payway in the .bashrc file? tks.