search

daswer123 / hallo-webui

Webui for Hallo: Hierarchical Audio-Driven Visual Synthesis for Portrait Image Animation
https://fudan-generative-vision.github.io/hallo/
59 stars 20 forks source link

Every step takes ~1 hour .. Too long #17

Open TomiTom1234 opened 3 weeks ago

TomiTom1234 commented 3 weeks ago

The default is on 40 steps. When I started generating, each step takes from 48m - 1 hour on my 3070 Ti (8gb VRAM).

Is it OK to make it only 3 - 5 steps ?