#Running the code requires the appropriate packages and dataset.
#The configurations to train the networks, as described in the paper, can be found in the 'configs' folder.
#The code also encompasses various alternatives that have been tested.
#Deeplip is an open-source library. We have integrated it directly since we have modified parts of the code.
#The CelebA folder contains the code for the CelebA dataset. This code has been developed on a separate branch. While the library and the deep learning models remain the same, the philosophy behind their deployment varies.

#for instance, to train the cat_vs_dog lip network run
python cat_vs_dog_train configs/cat_vs_dog_lip.yml

#to evalaute the metrix run 
cat_vs_dog_analysis [namle of your run]