ComfyUI Multi-Subject Workflows — Latent Couple Pose v1.1

Latent Couple Pose v1.1 is a ComfyUI workflow for generating multi-subject images with a separate prompt for each subject and OpenPose guidance; it can create a starter image for pose extraction or use a prepared pose image, then generate a final image.

By default, a basic text-to-image pass creates an image for the OpenPose Preprocessor to extract pose guidance; you can load a prepared pose image instead.

LoRA hooks apply LoRA models to image areas while sampling the whole image in one pass.

The starter image passes through an OpenPose Preprocessor to create pose guidance; the main output is a final image.

Download the Latent Couple Pose v1.1 model version entry; “Update Notification Push” is recreated with each release.

Vorschlag · nicht geprüft

To install required custom nodes, use ComfyUI Manager’s “Install Missing Custom Nodes” action; if any are still missing, check this version’s listed requirements.

Vorschlag · nicht geprüft

Version 1.1 lists these required custom nodes: ComfyUI-Custom-Scripts, ComfyUI Impact Pack, ComfyUI Impact Subpack, ComfyUI Inspire Pack, comfyui_controlnet_aux, and WAS Node Suite (Revised).

Named model files include sdxl_vae.safetensors, 4x-UltraSharp.pth, bbox/face_yolov8m.pt, bbox/hand_yolov8s.pt, controlnetxlCNXL_windsingaiTile10w.safetensors, ILXL CN OpenPose - windsingai.safetensors, ILXL [2D] Mistbloom - Rustyvivid.safetensors, [ZZZ] Jane Doe - okingjo.safet.

rgthree-comfy is recommended but not required; it helps mute groups quickly.

The Size Selector node in version 1.1 allows custom sizes.

For complex scenes, consider loading a pose prepared in an OpenPose editor; the default preprocessor may have trouble extracting pose data accurately.

Vorschlag · nicht geprüft

Check that the model filenames listed for this version match files in your ComfyUI model folders before running it.

Vorschlag · nicht geprüft