search

ajhamdi / MVTN

pytorch implementation of the ICCV'21 paper "MVTN: Multi-View Transformation Network for 3D Shape Recognition"
98 stars 9 forks source link

checkpoints #3

Closed whu-lee closed 2 years ago

whu-lee commented 2 years ago

Thank you very much for your reply. Sorry I have encountered another problem. What is the use of the weights under results/checkpoints/modelnet/? Could you provide the model weights for the best results?

ajhamdi commented 2 years ago

These are the weights of a ResNet18 pretrained on ModelNet40. This is required in the ViewGCN paper in order to do two-stage training where the first is single-view and the second is multi-view. The weights are provided to avoid doing the first stage, and they will be ignored if you end up doing the two stages by passing the flag --viewgcn_phase all