Hugging Face's logo

h94
/
IP-Adapter-FaceID

Text-to-Image
Diffusers
English
stable-diffusion
Model card Files
xet
 Community
73
multimodalart HF Staff commited on
Commit
55ef82c
·
1 Parent(s): e2f1fac

Update `image_encoder_path` to a public CLIP one

Browse files

Hi there, I'm proposing this PR to correct the IP Adapter Plus demo.
However, I'm not sure that is the correct CLIP model, as the correct CLIP model is not referred to anywhere

Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -149,7 +149,7 @@ from ip_adapter.ip_adapter_faceid import IPAdapterFaceIDPlus
149
  v2 = False
150
  base_model_path = "SG161222/Realistic_Vision_V4.0_noVAE"
151
  vae_model_path = "stabilityai/sd-vae-ft-mse"
152
- image_encoder_path = "h94/IP-Adapter/models/image_encoder"
153
  ip_ckpt = "ip-adapter-faceid-plus_sd15.bin" if not v2 else "ip-adapter-faceid-plusv2_sd15.bin"
154
  device = "cuda"
155
 
 
149
  v2 = False
150
  base_model_path = "SG161222/Realistic_Vision_V4.0_noVAE"
151
  vae_model_path = "stabilityai/sd-vae-ft-mse"
152
+ image_encoder_path = "laion/CLIP-ViT-H-14-laion2B-s32B-b79K"
153
  ip_ckpt = "ip-adapter-faceid-plus_sd15.bin" if not v2 else "ip-adapter-faceid-plusv2_sd15.bin"
154
  device = "cuda"
155