search

yuangan / EAT_code

Official code for ICCV 2023 paper: "Efficient Emotional Adaptation for Audio-Driven Talking-Head Generation".
Other
275 stars 31 forks source link

How can I replace the wav? #6

Closed skeletonNN closed 1 year ago

skeletonNN commented 1 year ago

This work is perfect, especially for the tooth generation. So when i replace the wav? What things i will do?

yuangan commented 1 year ago

Thank you for your attention. The preprocessing code is being prepared.

yuangan commented 1 year ago

Hi, I have updated the preprocessing code. Thank you for your support.

yuangan commented 1 year ago

Hi, I have made improvements to the demo code for replacing wav files. Simply replace the wav file named video1_name/video1_name.wav and the deep speech feature in video1_name/deepfeature32/video1_name.npy. Afterward, you can test your new wav by using the pose sequences from video1_name. For more details, please refer to this link. Enjoy~

G-force78 commented 1 year ago

This looks a neat update to styletalker/AVCT. Out of interests have you worked on the TED384 or tiktok models? How to test the different emotional expressions with custom video and image?

yuangan commented 1 year ago

Thank you for your attention. Our work is based on FoMM or Face-vid2vid and we have not tried TED384 or Tiktok models. As for testing custom videos and images, you can refer to our demo and preprocessing code.