https://responsivevoice.org/ # Spanish Latin America Female, Spanish female https://github.com/andabi/deep-voice-conversion https://github.com/michalfaber/keras_Realtime_Multi-Person_Pose_Estimation https://github.com/datitran/face2face-demo https://aescripts.com/auto-lip-sync/
gcloud auth login
gcloud compute --project "magggenta-176803" ssh --zone "us-central1-c" "gpu"
try head pose human pose split training data so we have data to train the gan that fixes shadows