search

ChenyangQiQi / FateZero

[ICCV 2023 Oral] "FateZero: Fusing Attentions for Zero-shot Text-based Video Editing"
http://fate-zero-edit.github.io/
MIT License
1.12k stars 108 forks source link

Questions: resolution and supported prompt #10

Closed bbecausereasonss closed 1 year ago

bbecausereasonss commented 1 year ago

Thanks for this!

  1. Is there a max resolution output on the video?
  2. Can we create any prompt/concept style through SD checkpoints or is it only through pretrained style YAML's you've provided?
ChenyangQiQi commented 1 year ago

Thanks for your interest!

  1. Due to the cost of memory and time, we only support 512x512 output resolution as that in stable diffusion
  2. I am not sure if I understand your question correctly. Basically, the applicable prompts/styles are beyond those in provided YAML. If your prompt/style works on unconditioned generation by stable diffusion, it is very likely to work in this editing framework. You may even try other finetuned checkpoints with the same network structure as stable diffusion