Keras version of SyncNet, by Joon Son Chung and Andrew Zisserman.
SyncNet paper: "Out of time: automated lip sync in the wild"
VGG webpage: VGG - SyncNet
-
Libraries required by Python (I used Python 3) are mentioned in the requirements.txt file.
-
Pre-trained weights, to be placed in the syncnet-weights directory. Instructions to download and place the files are available in the readme file inside the syncnet-weights directory
-
SyncNet takes input images of size (112, 112, 5).
-
These input images have pixel values between 0 and 255! DON'T rescale image values to [0, 1], keep them in [0, 255].