README.md

metadata

license: mit
tags:
  - comfyui
  - stable-diffusion-xl
  - controlnet
  - ipadapter
  - clip-vision
  - vae
  - upscale
  - inpaint

ComfyUI Model Collection (SDXL)

A curated collection of models for ComfyUI workflows, focused on SDXL-based image generation. Includes ControlNet, IPAdapter, CLIP Vision, VAE, inpainting, upscaling, and detection models.

---

📁 Repository Structure

├── clip_vision/          # CLIP Vision encoders for IPAdapter
├── controlnet/SDXL/      # ControlNet models for SDXL
├── inpaint/              # Inpainting models
├── ipadapter/SDXL/       # IPAdapter models for SDXL
├── sams/                 # SAM (Segment Anything) models
├── ultralytics/
│   ├── bbox/             # Bounding box detection models (YOLO)
│   └── segm/             # Segmentation models (YOLO)
├── upscale_models/       # Image upscaling models
└── vae/SDXL/             # VAE models for SDXL

---

🧩 Models Included

CLIP Vision

File	Description
clip-vision_vit-h.safetensors	ViT-H CLIP Vision encoder
CLIP-ViT-bigG-14-laion2B-39B-b160k.safetensors	BigG CLIP Vision (large)
CLIP-ViT-H-14-laion2B-s32B-b79K.safetensors	ViT-H CLIP Vision (laion2B)
clip_vision_g.safetensors	CLIP Vision G encoder

ControlNet (SDXL)

File	Description
controlnet-union-sdxl-1.0-promax.safetensors	Union ControlNet (multi-purpose)
control-lora-canny-rank256.safetensors	Canny edge detection
control-lora-depth-rank256.safetensors	Depth map
noob-sdxl-controlnet-*.fp16.safetensors	NoobAI SDXL ControlNet series (canny, depth, lineart, normal, scribble, softedge, tile)
noobaiXLControlnet_openposeModel.safetensors	OpenPose
IllustriousXL_openpose.safetensors	OpenPose (IllustriousXL)
OpenPoseXL2.safetensors	OpenPose XL v2
AnytestV4.safetensors	Anytest v4 base model
fabricatedXL_v70.safetensors	FabricatedXL v7.0
waiIllustriousSDXL_v160.safetensors	Wai IllustriousSDXL v1.6
noobaiInpainting_v10.safetensors	NoobAI Inpainting

IPAdapter (SDXL)

File	Description
ip-adapter-plus_sdxl_vit-h.safetensors	IPAdapter Plus
ip-adapter-faceid-plusv2_sdxl.bin	FaceID Plus v2
ip-adapter-faceid-portrait_sdxl.bin	FaceID Portrait
ip-adapter-faceid-portrait_sdxl_unnorm.bin	FaceID Portrait (unnormalized)
ip-adapter-faceid_sdxl.bin	FaceID base
noobIPAMARK1_mark1.safetensors	NoobAI IPAdapter Mark1

Inpaint

File	Description
inpaint_v26.fooocus.patch	Fooocus inpaint patch v2.6
fooocus_inpaint_head.pth	Fooocus inpaint head
MAT_Places512_G_fp16.safetensors	MAT inpainting model

SAM (Segment Anything)

File	Description
sam_vit_b_01ec64.pth	SAM ViT-B

Ultralytics — BBox Detection

File	Description
face_yolov8m.pt	Face detection (YOLOv8m)
face_yolov9c.pt	Face detection (YOLOv9c)
hand_yolov9c.pt	Hand detection (YOLOv9c)
femaleBodyDetection_yolo26.pt	Body detection
Eyeful_v2-Paired.pt	Eye detection

Ultralytics — Segmentation

File	Description
face_yolov8m-seg_60.pt	Face segmentation
person_yolov8m-seg.pt	Person segmentation
femaleBodyDetection_typea.pt	Body segmentation
yolo11m-seg.pt	General segmentation (YOLO11m)

Upscale Models

File	Description
4x_foolhardy_Remacri.pth	4x upscale (Remacri)
4x-AnimeSharp.pth	4x anime upscale
4x_NMKD-Superscale-SP_178000_G.pth	4x NMKD Superscale
2x_APISR_RRDB_GAN_generator.pth	2x APISR upscale
4x_APISR_GRL_GAN_generator.pth	4x APISR GRL upscale
OmniSR_X2/X3/X4_DIV2K.safetensors	OmniSR upscale (2x/3x/4x)

VAE (SDXL)

File	Description
sdxl_vae.safetensors	Standard SDXL VAE
illustriousXLV10_v10.safetensors	IllustriousXL VAE

---

🚀 How to Use with ComfyUI

Place models in the corresponding ComfyUI folders:

ComfyUI/models/
├── clip_vision/       ← clip_vision/*
├── controlnet/        ← controlnet/SDXL/*
├── inpaint/           ← inpaint/*
├── ipadapter/         ← ipadapter/SDXL/*
├── sams/              ← sams/*
├── ultralytics/       ← ultralytics/bbox/* and ultralytics/segm/*
├── upscale_models/    ← upscale_models/*
└── vae/               ← vae/SDXL/*

Or download with huggingface-cli:

huggingface-cli download danaidesign/comfyui --local-dir ./ComfyUI/models