Last version of ComfyUI break PuLID :

For a quick fix, find the file comfy/ldm/flux/model.py in your installation (this is the file to which the error message refers) and replace this line (also mentioned in the error message):

        out = self.forward_orig(img, img_ids, context, txt_ids, timestep, y, guidance, control, transformer_options, attn_mask=kwargs.get("attention_mask", None))

with

        out = self.forward_orig(img, img_ids, context, txt_ids, timestep, y, guidance, control, transformer_options)

Restart ComfyUI and the error should no longer appear (and apparently the generation still works as expected).

UmeAiRT PuLID simple workflow

Today I'd like to share with you a slightly more specific workflow.

PuLID allows you to create images of a person from a single photo!

I've written a complete guide to its use: Step-by-Step Guide Series: ComfyUI - PuLID | Civitai

How to install

Automatic

The prerequisites are a little more complex than usual, but the easiest way to install the necessary nodes is to use my "add missing node script" or "ComfyUI auto installer".

Manual

For manual instruction wait my guide or just look here :
Gourieff/comfyui-reactor-node: Fast and Simple Face Swap Extension Node for ComfyUI

描述:

New simple version, civitai metadata not include.

训练词语: