A pytorch implemention of MoCoGAN
102
stars
21
forks
source link
MoCoGAN: Decomposing Motion and Content for Video Generation
Description
Implemention
- you need to place datasets([run, walk, skip, ......]) here under 'raw_data' dir.
python resize.py
: resize your raw data as 96x96 and place them under resized_data dir.
python train.py
: train MoCoGAN.
Results
Requirements
- python 3.5
- pytorch 0.2
- numpy
- skvideo
- ffmgeg (for preprocessing and saving videos)
References