CNN + LSTM for word lip reading using the GRID dataset.
cd data ./get_data.sh
cd vgg16 python get_vgg16.py