Does anyone have the experience or resources of fine-tunning a pretrained optical flownet as a sub-learning module in a PyTorch learning model? I found a couple of implementation of the Pytorch Flownet pretrained model. However, as a novice, I want to know the best-practice or the convenient way to incorporate the optical flow learning model as a sub-model in a PyTorch learning model.
Any type of help is really appreciated. Thanks in advance.