search

shchur / overlapping-community-detection

Implementation of "Overlapping Community Detection with Graph Neural Networks"
http://www.daml.in.tum.de/nocd
MIT License
157 stars 45 forks source link

RuntimeError #5

Open atheeraa opened 3 years ago

atheeraa commented 3 years ago

Hi, I converted the notebook into py but I got this error when running the code:

RuntimeError: An attempt has been made to start a new process before the current process has finished its bootstrapping phase.

Do you have any idea of what's causing it?

shchur commented 3 years ago

Most likely this is some problem with multiprocessing on Windows. Setting num_workers=0 in the line sampler = nocd.sampler.get_edge_sampler(A, batch_size, batch_size, num_workers=5) might help.

Bisma1 commented 2 years ago

hi, can you please help me , i am not getting any output even after installing every said module in the code it shows "torch not compiled with cuda " i am running this notebook code on VScode . thankyou.

shchur commented 2 years ago

This means your computer doesn't have a CUDA-compatible GPU. You can try running this code on Google Colab, they provide free GPUs there.

Bisma1 commented 2 years ago

Hi , can you please define me how to give an edge list as input to this training model . I mean to say how can i convert my edge list into the same DATASET format which is given in this training model ..Thankyou