A redundant layer in PyTorch implementation

NVlabs / PWC-Net

PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume, CVPR 2018 (Oral)

Other

1.63k stars 357 forks source link

A redundant layer in PyTorch implementation #110

Open littlespray opened 4 years ago

littlespray commented 4 years ago

Hi,

I find that in PyTorch version , line 122, there is a deconv layer that has been defined but remains unused. If it is a typo, May I delete it? Otherwise, I have to use find_parameters_`unused=Truewhen using nn.parallel.DistributedDataParallel.

Thank you very much!

littlespray commented 4 years ago

The attribute is find_unused_parameters, sorry that I wasn't clear.