search

levihsu / OOTDiffusion

Official implementation of OOTDiffusion: Outfitting Fusion based Latent Diffusion for Controllable Virtual Try-on
Other
5.53k stars 809 forks source link

garm img use clip image encode #126

Closed kisstea closed 7 months ago

kisstea commented 7 months ago

I don't find the code , to encode garment image use clip image encode as the input of gram unet in inference

levihsu commented 7 months ago

https://github.com/levihsu/OOTDiffusion/blob/344112ad1c03c2af1cf7a1f07d689b18af4c175a/ootd/inference_ootd.py#L109

See prompt_embeds here

kisstea commented 7 months ago

ok, thx