dataset=$1 # Download the videos. python videos_download.py --input_list data_list/${dataset}_video_ids.txt --output_dir ${dataset}/raw_videos # Split the videos into 1-min chunks. ./videos_split.sh ${dataset}/raw_videos ${dataset}/1min_clips # Extract the talking head clips. python videos_crop.py --input_dir ${dataset}/1min_clips/ --output_dir ${dataset}/cropped_clips --clip_info_file data_list/${dataset}_video_tubes.txt