text2video_audio.sh

export LANG="en_US.UTF-8"
# desktop
# source venv_vid2vid/bin/activate 

echo "$1"
echo $2

rm -f input/$2/"$1".txt
echo "$1" >> input/$2/"$1".txt

# eng
cd aligner
python align_english_real_audio.py "$1" $2 
# python align_english_states.py "$1" $2 

cd ..

# ffmpeg -i input_audio_real/$2/Shehadyour.wav input_audio_real/$2/Shehadyour.mp3

# chn
# python pinyin_timestamping.py $1 $2

# generate video
rm -f ../vid2vid/datasets/$2/test_openpose/tmp/*
rm -f ../vid2vid/datasets/$2/test_img/tmp/*

rm -f ../vid2vid/datasets/$2/test_openpose/tmp_smooth/*
rm -f ../vid2vid/datasets/$2/test_img/tmp_smooth/*

## dataset and smooth method 
python interp_landmarks_motion_phoneme_VidTIMIT_smooth.py "$1" $2
# python interp_landmarks_motion_phoneme_VidTIMIT.py "$1" $2
# python interp_landmarks_motion_phoneme_dict.py "$1" $2
# python interp_landmarks_motion.py $1 $2
# python interp_landmarks_motion_phoneme.py "$1" $2

cd ../vid2vid/

rm -f results/$2/test_latest/tmp/*
rm -f results/$2/test_latest/tmp_smooth/*

CUDA_VISIBLE_DEVICES=1 python test.py --name $2 --dataroot datasets/$2 --dataset_mode pose --input_nc 3 --resize_or_crop scaleHeight --loadSize 512 --openpose_only --how_many 1200 --no_first_img --random_drop_prob 0

python image2video_real_audio_text2video.py "$1" $2