Simple SDXL face workflow for newbies like me-- With IPA plus face, Reactor, UltSDUpscale
5.0
0 reviews
Description
-----
A lot of the workflows here on openart.ai, even the ones labeled as "Launch on Cloud", don't work due to different user environments.
But this was all built here, so it shoud load up with no extra work or hunting down models or custom nodes.
**If you run into problems, remember to always update your custom nodes and models upon first loading the workflow**
-----
I built this because I cannot get any of the IPA+FaceID workflows hosted on Openart to work!
Those FaceID models an nodes and models are quite new, so once they are rolled out and working may render this workflow redundant or outdated. In the mean time....
This is a simple workflow for character design and faceswap-- Adding an IP Plus face adapter into the workflow before the Reactor model helps maintain face-shapes and overall consistency.
Let me know if you have any questions or suggestions.
Node Diagram
Discussion
laion/CLIP-ViT-H-14-laion2B-s32B-b79K ?
where i can download this model? i have reached the huigginface, folder have that name. but there is no safetensors with that name
I am very new to this, and have only been using OpenArt's web version of Comfy, but I think that alot of the Clip vision model safetensors are simply named "model.safetensors. You download the model and rename it to anything you want. That appears to be the case here. On the Huggingface page for the clip model here: https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K/tree/main I believe you want the "model.safetensors" file listed in the "files and versions: tab for the model:
https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79K/blob/main/model.safetensors
Just rename it to something that makes sense to you and refernce that in the comfy workflow.
The IPA Plus SDXL models are a bit funny in that you use a SD1.5 clip vision model with the SDXL IPAs. I think that trips up a lot of people at first, it certainly did me.
I hope this helps. Cheers
thank you for the reply, but i have already download the model.safetensors, but i still get the error,
What does the error say?
thanks for sharing, it works so far. I am also new to this, and it is hard to find workflows that actually works.
Thanks! I just started by finding workflows that were reliably running out-of-the-box and modified them as I needed.
Version 2 of this workflow is coming soon, with the addition of openpose ControlNet, tweaks, and more options.
I look forward to updates from you.
Node Details
Primitive Nodes (17)
IPAdapterApply (1)
Note (7)
Reroute (9)
Custom Nodes (20)
ComfyUI
- CheckpointLoaderSimple (1)
- CLIPVisionLoader (1)
- UpscaleModelLoader (1)
- PreviewImage (3)
- CLIPTextEncodeSDXL (2)
- VAEDecode (1)
- KSampler (1)
- EmptyLatentImage (1)
- LoadImage (2)
- SaveImage (1)
- PrepImageForClipVision (1)
- IPAdapterModelLoader (1)
- CropFace (1)
- ReActorFaceSwap (1)
- UltimateSDUpscale (1)
- Bus Node (1)
Model Details
Checkpoints (1)
juggernautXL_v8Rundiffusion.safetensors
LoRAs (0)