| 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 |