Why can Nvidia RTX Voice run on my GTX 1080, but Nvidia Maxine requires tensor cores and won't work?

AI & Data Science Deep Learning (Training & Inference) Maxine
1

I have a GTX 1080 graphics card and would like to experiment with Nvidia’s audio-denoising capabilities as a developer. While I can run Nvidia RTX Voice without any issues, it only works with a microphone as input, not with audio files. Nvidia Maxine seems to provide the functionality I need for file-based experimentation, but it requires tensor cores, which are not compatible with my 1080 card.

How is Nvidia RTX Voice able to run on a GTX 1080, while Nvidia Maxine can’t? What are my options for using Nvidia’s audio-denoising features with a GTX 1080?

2

RTX Voice is not part of the Maxine software development kit and we do not support Maxwell and older GPUs. To leverage the Maxine SDK you will need at least a Turing generation GPU. RTX Voice’s model was built specifically for GTX and that model is not maintained by the Maxine team.