wwMark / meshgraphnets

Rewrite deepmind/meshgraphnets into pytorch
72 stars 22 forks source link

a bug in the function find_influential_nodes #11

Open oshrihalimi1 opened 2 years ago

oshrihalimi1 commented 2 years ago

I believe this line meant to sum the number of neighbors and not the neighbor indices: neighbors = list(map(sum, indices)) I think it should be: neighbors = list(map(len, indices))

Did the ripple training worked for you? Which method had the best performance with cloth simulation?