Inference time

[skunkwerk]

Hi, thanks for publishing this work!
I couldn’t find any mention in your paper of the inference time for generating a new model from an image / text. Do you know how long it takes, and on what GPU you tested on?

[menyifang]

For inference, it runs in real-time for generative module and 30s for two optimize modules. When working from image/text to the full scheme, inversion and some mesh operations should be considered. So it takes around 3 minutes in total on a 3090 GPU.