Model
Trends
.ai
Model
Trends
.ai
die besten Open-Source-KI-Modelle

IPAdapter (IP-Adapter Plus)-Node in ComfyUI

The IPAdapter nodes feed a reference picture into the model as an "image prompt", transferring style, composition or a face without describing them in words; IPAdapter Unified Loader plus IPAdapter Advanced are the usual pair.
Node-Paket: ComfyUI_IPAdapter_plus
Kategorie: ControlNet und Steuerung
Diese Seite ist noch nicht übersetzt und wird deshalb auf Englisch angezeigt.

Was IPAdapter (IP-Adapter Plus) macht

IP-Adapter is an image-conditioning method: a CLIP Vision encoder reads a reference image and a small adapter injects that embedding into the attention layers of the diffusion model. The ComfyUI_IPAdapter_plus pack (by cubiq/matteo) wraps it in a few nodes. IPAdapter Unified Loader picks the right adapter and CLIP Vision model for your checkpoint by preset (LIGHT, STANDARD, PLUS, PLUS FACE, FULL FACE, VIT-G). IPAdapter Advanced applies it with weight, weight_type (linear, style transfer, composition, strong style transfer), start/end and an optional attention mask.
Typical uses: copy the style of a painting onto a new subject, keep a character consistent across renders, or transfer a composition. With weight_type "style transfer" and a mask you can restrict the influence to part of the image. FaceID variants add an InsightFace embedding for stronger identity; they need the insightface dependency.

Eingänge

Name
Typ
Bedeutung
model
MODEL
From the checkpoint or LoRA loader.
ipadapter
IPADAPTER
From IPAdapter Unified Loader (or IPAdapter Model Loader).
image
IMAGE
The reference picture (square 224 crop is taken; use Prep Image for IPAdapter).
weight
FLOAT
0-1 typical; 0.7-0.8 for style, 0.5 for subtle influence.
weight_type
COMBO
linear, ease in/out, style transfer, composition, strong style transfer and more.
start_at / end_at
FLOAT
Sampling range where the adapter is active.
attn_mask
MASK
Optional; restricts the influence to a region.

Ausgänge

Name
Typ
Bedeutung
MODEL
MODEL
The patched model; connect it to KSampler.
AD
Spar dir die Einrichtung: ComfyUI-Workflows vorinstalliert in der Cloud
BitVector führt fertige ComfyUI-Workflows auf eigenen GPUs aus. Keine Installation, keine fehlenden Nodes, keine roten Kästen. Auf dem Handy oder Laptop öffnen und in einer Minute generieren.

So verwendest du IPAdapter (IP-Adapter Plus)

  1. Install ComfyUI_IPAdapter_plus from ComfyUI Manager and download the ipadapter and clip_vision files it lists (models/ipadapter, models/clip_vision).
  2. Add IPAdapter Unified Loader after Load Checkpoint, preset PLUS (high strength).
  3. Add IPAdapter Advanced; connect model, ipadapter and the reference IMAGE.
  4. Set weight 0.8, weight_type "style transfer" for style or "linear" for a general reference.
  5. Connect the MODEL output to KSampler and render with your prompt.

Einstellungen und Tipps

  • Prep Image for IPAdapter crops and sharpens the reference to the 224 px square the encoder sees.
  • Several references: use IPAdapter Batch or the Encoder + Combine Embeds nodes.
  • Lower weight and end_at 0.7 when the prompt is ignored.
  • PLUS FACE and FaceID work on portraits; for body consistency use PLUS with a full-body reference.
  • Unified Loader needs exact file names (see the pack README) to auto-detect models.

Fehlerbehebung bei IPAdapter (IP-Adapter Plus)

ClipVision model not found / IPAdapter model not found

Warum es passiert
The Unified Loader looks for specific file names in models/clip_vision and models/ipadapter and did not find them.

So behebst du es
Download the files named in the pack README (for example CLIP-ViT-H-14-laion2B-s32B-b79K.safetensors and ip-adapter-plus_sdxl_vit-h.safetensors) and keep those exact names.

Error: insightface model is required for FaceID models

Warum es passiert
FaceID presets need the insightface Python package and the buffalo_l detector.

So behebst du es
Install insightface and onnxruntime in the ComfyUI Python environment (prebuilt wheels on Windows), or use PLUS FACE instead.

The prompt is ignored, output is a copy of the reference

Warum es passiert
weight too high with the linear type.

So behebst du es
Weight 0.5-0.7, end_at 0.6-0.8, or weight_type style transfer.

Error: Expected size 1280 but got 1024 (or similar)

Warum es passiert
The CLIP Vision model does not match the adapter (ViT-H adapters need the ViT-H encoder; VIT-G needs bigG).

So behebst du es
Use the Unified Loader presets, or pair the files as listed in the README.

AD
Diesen Workflow vom Handy aus starten
Jeder Workflow auf dieser Seite ist auf BitVector vorinstalliert, mit Modellen und Custom Nodes an Ort und Stelle. Einen auswählen, Prompt eingeben, fertig. Null Einrichtung, nichts herunterladen.

Fragen zu IPAdapter (IP-Adapter Plus)

IP-Adapter or ControlNet?

ControlNet copies structure (edges, pose) exactly; IP-Adapter copies look and feel (style, identity, mood) loosely. They combine well.

Does it work with Flux?

Yes, with the XLabs or InstantX Flux IP-Adapter models and the Flux-specific loader nodes in the pack.

Can I use several references?

Yes: batch the images (Image Batch) or combine embeddings with IPAdapter Encoder and Combine Embeds.

Verwandte Nodes

Apply ControlNet
ComfyUI core
Apply ControlNet attaches a guide image and a loaded ControlNet to the positive and negative CONDITIONING with strength and start/end percentages, so the sampler follows edges, depth or pose from the guide.
Load Image
ComfyUI core
Load Image reads a picture from the ComfyUI input folder (or an upload) and outputs an IMAGE tensor plus a MASK taken from the alpha channel, the starting point for img2img, inpainting, ControlNet and IP-Adapter graphs.
Load Checkpoint
ComfyUI core
Load Checkpoint (CheckpointLoaderSimple) opens a .safetensors or .ckpt model file and hands out the three parts every workflow needs: the diffusion MODEL, the CLIP text encoder and the VAE.
KSampler
ComfyUI core
KSampler runs the denoising loop: it takes the model, prompts and a latent and produces the finished latent image, controlled by seed, steps, cfg, sampler, scheduler and denoise.
FaceDetailer (Impact Pack)
ComfyUI-Impact-Pack
FaceDetailer finds faces (or hands, eyes) with a YOLO detector, crops each one, re-renders it at full resolution with your model and prompt, and pastes it back, fixing the small, mushy faces in wide shots.
AD
Keine Lust mehr, Nodes zu reparieren? Lass es die Cloud machen
BitVector hält hunderte ComfyUI-Workflows installiert, aktuell und getestet auf schnellen Cloud-GPUs. Kein Python, keine CUDA-Fehler, keine VRAM-Grenzen. Läuft in jedem Browser.

Weitere ComfyUI-Nodes

CLIP Text Encode (Prompt)
ComfyUI core
CLIP Text Encode turns a text prompt into CONDITIONING using the model text encoder. One node holds the positive prompt, a second one the negative prompt.
ComfyUI Manager
ComfyUI-Manager
ComfyUI Manager is the extension that installs, updates and fixes custom node packs and models from inside the interface, resolves missing nodes in imported workflows and snapshots your setup.
Save Image
ComfyUI core
Save Image writes the IMAGE tensor to ComfyUI/output as a PNG, with the whole workflow embedded in the file metadata so the picture can be dragged back into ComfyUI to restore the graph.
VAE Decode
ComfyUI core
VAE Decode converts the sampled LATENT into a pixel IMAGE with the VAE; VAE Decode (Tiled) does the same in tiles for very large images.
Empty Latent Image
ComfyUI core
Empty Latent Image creates the blank latent canvas (width, height, batch_size) that text-to-image sampling starts from; dimensions must be multiples of 8 and match the model family.
Load LoRA
ComfyUI core
Load LoRA (LoraLoader) applies a LoRA file to the MODEL and CLIP with separate strengths, so a style, character or concept can be added to any checkpoint without merging.