FlowNetPytorch

Pytorch implementation of FlowNet by Dosovitskiy et al.

This repository is a torch implementation of FlowNet, by Alexey Dosovitskiy et al. in PyTorch. This code is mainly inspired from here

Deformable Convolutions as seen in this paper, code inspired by

Following neural network models are currently provided :

Prerequisite

pytorch >= 0.4.1
tensorboard-pytorch
tensorboardX >= 1.4
spatial-correlation-sampler>=0.0.8
imageio
argparse

python main.py /path/to/training_set/ -b8 -j8 -a flownets -s path/to/split/file

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
Pytorch-Correlation-extension		Pytorch-Correlation-extension
__pycache__		__pycache__
datasets		datasets
deform_conv_pytorch		deform_conv_pytorch
models		models
README.md		README.md
Sintel_train_val.txt		Sintel_train_val.txt
flow_lib.py		flow_lib.py
flow_transforms.py		flow_transforms.py
main.py		main.py
makeitrun.txt		makeitrun.txt
multiple_inference.py		multiple_inference.py
multiscaleloss.py		multiscaleloss.py
requirements.txt		requirements.txt
run_inference.py		run_inference.py