ashawkey / RAD-NeRF

Real-time Neural Radiance Talking Portrait Synthesis via Audio-spatial Decomposition
MIT License
863 stars 151 forks source link

How to make it work realtime? #58

Open Luckygyana opened 1 year ago

iboyles commented 11 months ago

I second this what hardware is required for real time? On colab for running inference after having to run asr.py on any audio before hand makes it 4 to 5 times the length of real time