반응형
self.model.Traj_Dec.decoder_mlp[0].weight.data = checkpoint['scratch_state_dict']['Traj_Dec.decoder_mlp.0.weight']
self.model.Traj_Dec.decoder_mlp[0].bias.data = checkpoint['scratch_state_dict']['Traj_Dec.decoder_mlp.0.bias']Deep Learning
[Pytorch] Assign (loaded) pre-trained model params to a specific layer of a new model
'Deep Learning' Related Articles
-
How to make novel ideas?
- [Pytorch] Stop updating params of a layer of a model during back-propagation
- D. Choi et al., "Regularizing neural networks for future trajectory prediction via inverse reinforcement learning framework," IET Computer Vision, In Press.
- [Pytorch] 2D Conv using F.conv2d (2d conv with a pre-defined filter)
