Deep3DFaceRecon_pytorch icon indicating copy to clipboard operation
Deep3DFaceRecon_pytorch copied to clipboard

No module named 'torch'

Open davidSirovsky opened this issue 2 years ago • 1 comments

I tried to launch a project in the Google Collab service. I did everything as indicated in the instructions, but when I try to run the script

get reconstruction results of example images

python test.py --name=<model_name> --epoch=20 --img_folder=./datasets/examples

an error occurs - No module named 'torch'.

Even if I re-install torch, then it will require other libraries that were also already installed before. Even if I install all the libraries manually again, after that another error appears, which says that the CUDA is missing, although CUDA (GPU) is activated in the Collab settings

After installing nvdiffrast, the following warning appeared: WARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv Don't know if this is related to the main problem

davidSirovsky avatar Apr 21 '22 13:04 davidSirovsky

(Commenting because I had this same issue) You skipped the source activate deep3d_pytorch step. On my machine, I had to use conda activate deep3d_pytorch. As I understand it, that basically tells Anaconda to actually use everything it downloaded and installed in this terminal.

I needed to redo the pip install . step from nvdiffrast for the same reason - it hadn't been included as part of that environment.

DrCyanide avatar Feb 14 '23 08:02 DrCyanide