weight in GTConv is alwasy the same

seongjunyun / Graph_Transformer_Networks

Graph Transformer Networks (Authors' PyTorch implementation for the NeurIPS 19 paper)

960 stars 179 forks source link

weight in GTConv is alwasy the same #21

Closed LuciusMos closed 4 years ago

LuciusMos commented 4 years ago

When I print weight in GTConv (Actually I print Ws in GTN), I find it is always [[0.2, 0.2, 0.2, 0.2, 0.2], [0.2, 0.2, 0.2, 0.2, 0.2]], seems like it never updates. Why would this happen?

LuciusMos commented 4 years ago

I'm sorry that there is already a solution. Click Here for details. Thanks @changym3 so much! BTW, the torch-geometric version cannot be 1.4.3. I met another problem and it is solved after I reinstalled torch-geometric==1.4.2