Aller au contenu
WorkflowPrésentéExterne
Télécharger

Qwen Image Edit 2511 Factory Version 2 (Regular Use/BG Expand/Inpaint/Pose Changer, Refiners, SeedVR2 Upscaler) — v2.0

[Improvement] For Case 1 (Standard Use), we've enabled the use of multiple images (turn it OFF if you don't use it). This expands the possibilities, such as changing background images.

Sourcecivitai.com/models/2264596
TâcheMultipleCapacitésÉdition d'image · Image en image · Texte en image · Mise à l'échelleFamilleQwen-ImageVersionRelease 2551814 — QwenImageEdit2511_Factory2.jsonEstimé VRAMNot yet characterizedUnratedClasse

Profil d'exécution

Accès
Flux de travail externe
Statut
Présenté
Version
QwenImageEdit2511_Factory2.json
Source
Civitai
Mis à jour
3 nov. 2026

Description de la source

Release Note

(v1.0 -> v1.1)

[Improvement] For Case 1 (Standard Use), we've enabled the use of multiple images (turn it OFF if you don't use it). This expands the possibilities, such as changing background images.

[Improvement] Added Refiner (SDXL). If you created (image 1) using SDXL, this should help make the generated image quality closer to the original.

What you need

Model files
9.5 GiB known subtotal

Remaining hardware or download-size estimates are not available yet.

Exigences

Modèles

14

  • DaeXL.safetensors

    Non résolu

    Checkpoint · Unknown

  • ema_vae_fp16.safetensors

    Non résoluOptionnel

    VAE · Unknown

  • qwen_2.5_vl_7b_fp8_scaled.safetensors

    Fichier indiqué

    Text encoder · 8.74 GB · Hugging Face · Comfy-Org/HunyuanVideo_1.5_repackaged

  • qwen_image_edit_2511_bf16.safetensors

    Non résolu

    Diffusion model · Unknown

  • qwen_image_vae.safetensors

    Dépôt identifié

    VAE · Hugging Face · Comfy-Org/Qwen-Image_ComfyUI

  • Qwen-Image-Edit-2511-Lightning-4steps-V1.0-bf16.safetensors

    Trouvé

    LoRA · 810 MB · SAFETENSOR · BF16 (tensor header) · Hugging Face · lightx2v/Qwen-Image-Edit-2511-Lightning

  • sd_xl_refiner_1.0_0.9vae.safetensors

    Non résolu

    VAE · Unknown

  • sdxl-boldline.safetensors

    Non résolu

    Checkpoint · Unknown

  • sdxl-flat.safetensors

    Non résolu

    Checkpoint · Unknown

  • seedvr2_ema_7b_sharp_fp16.safetensors

    TrouvéOptionnel

    Checkpoint · 15.3 GB · SAFETENSORS · Unknown

  • Unresolved model selection: LayerMask: LoadSegmentAnythingModels

    Non résoluOptionnel

    Unknown

  • usako_qwen.safetensors

    Non résoluOptionnel

    LoRA · Unknown

  • vxpXLLCM_vxpXLV10.safetensors

    Non résolu

    Checkpoint · Unknown

  • x1_ITF_SkinDiffDetail_Lite_v1.pth

    Non résoluOptionnel

    Upscaler · Unknown

Packs de nœuds

19

  • ComfyUI_Comfyroll_CustomNodes

    Trouvé

    Pack de nœud · Registry

  • comfyui_controlnet_aux

    PossibleOptionnel

    Pack de nœud · Registry

  • comfyui_layerstyle

    Trouvé

    Pack de nœud · Registry

  • ComfyUI_LayerStyle_Advance

    TrouvéOptionnel

    Pack de nœud · Registry

  • comfyui-custom-scripts

    Trouvé

    Pack de nœud · Registry

  • comfyui-easy-use

    TrouvéOptionnel

    Pack de nœud · Registry

  • comfyui-inpaint-nodes

    TrouvéOptionnel

    Pack de nœud · Registry

  • comfyui-kjnodes

    Trouvé

    Pack de nœud · Registry

  • comfyui-logicutils

    TrouvéOptionnel

    Pack de nœud · Registry

  • comfyui-post-processing-nodes

    TrouvéOptionnel

    Pack de nœud · Registry

  • ComfyUI-QwenVL

    PossibleOptionnel

    Pack de nœud · Registry

  • ComfyUI-tbox

    Trouvé

    Pack de nœud · Registry

  • Fast Groups Muter (rgthree)

    Non résolu

    Pack de nœud · Unknown

  • GetNode

    Non résolu

    Pack de nœud · Unknown

  • masquerade-nodes-comfyui

    TrouvéOptionnel

    Pack de nœud · Registry

  • rgthree-comfy

    Trouvé

    Pack de nœud · Registry

  • seedvr2_videoupscaler

    Non résoluOptionnel

    Pack de nœud · Registry

  • SetNode

    Non résolu

    Pack de nœud · Unknown

  • was-ns

    Non résolu

    Pack de nœud · Registry

Commentaire généré par l’IA

Explication générée par l’IA à partir des détails de la source et de la configuration. Les suggestions sont clairement signalées.

Objectif

  • Use Qwen Image Edit 2511 Factory Version 2 (v2.0) to edit source images in standard, background-extension, inpainting, or pose-change mode; inputs vary by mode and include a source image, prompt, resolution, mask, or pose reference, and the output is an edited image.

Flux

  • The workflow includes PreviewImage and SaveImage nodes for previewing and saving an image.

  • A simple Refiner and an Upscaler are optional; the Refiner is described as helping keep an image created with SDXL closer to its original.

Entrées et sorties

  • For standard edits, upload a source image to image 1 and enter a prompt; this mode also supports multiple images, including for changing backgrounds.

  • For background extension, upload a source image to image 1 and set the resolution; for inpainting, upload a source image to image 1, edit the mask, and enter a prompt.

  • For pose changes, upload the original image to image 1 and a pose-reference image to image 2; an HD pose image is preferred.

Dépendances

  • The workflow lists these node dependencies: ComfyUI_Comfyroll_CustomNodes, comfyui_controlnet_aux, comfyui_layerstyle, ComfyUI_LayerStyle_Advance, comfyui-custom-scripts, comfyui-easy-use, comfyui-inpaint-nodes, comfyui-kjnodes, comfyui-logicutils, comfyui-post-processing-nodes,

  • Other listed node dependencies are ComfyUI-QwenVL, ComfyUI-tbox, Fast Groups Muter (rgthree), GetNode, masquerade-nodes-comfyui, rgthree-comfy, seedvr2_videoupscaler, SetNode, and was-ns.

  • Listed model files include DaeXL.safetensors, ema_vae_fp16.safetensors, qwen_2.5_vl_7b_fp8_scaled.safetensors, qwen_image_edit_2511_bf16.safetensors, qwen_image_vae.safetensors, Qwen-Image-Edit-2511-Lightning-4steps-V1.0-bf16.safetensors, sd_xl_refiner_1.0_0.9vae.safetensors, sdx

Configuration

  • Tip: In standard mode, turn off the multiple-image option when you are not using multiple images.

    Suggestion · non vérifié

  • Tip: For pose changes, use DW or Depth for ControlNet.

    Suggestion · non vérifié

  • The workflow note lists reference settings of Qwen: 40 steps/CFG 4.0, Comfy: 20 steps/CFG 4.0, and Lightning LoRA: 4 steps/CFG 1.0.

Précautions

  • The listing does not name a model file for LayerMask: LoadSegmentAnythingModels; check which model your setup needs for that node.

    Suggestion

  • A usage license is not specified in the listing; check the source terms before use.

    Suggestion

Faut-il modifier cela ?

Connectez-vous pour envoyer une demande de modification.

Sources

1 source
CivitaiPrincipalVersion 2551814
civitai.com/models/2264596

Vérifié 3 nov.

Workflow instructions

Author notes from QwenImageEdit2511_Factory2.json · selected release

You can test and find the best setting by yourself. The following table is for reference. | | Qwen | Comfy | lightning LoRA | |--------|---------|------------|---------------------------| | Steps | 40 | 20 | 4 | | CFG | 4.0 | 4.0 | 1.0 | By default, we use 20 steps as we just don't want it to take too long. Try 40 if you want a better result, but it will take longer.

Author notes from QwenImageEdit2511_Factory2.json · selected release

## When using this mode, you must first edit the mask (image 1) to define the Inpaint area. Otherwise, an error will occur.

Author notes from QwenImageEdit2511_Factory2.json · selected release

## If your output doesn't work well (especially confirmed with anime-style images), it might be better to describe the character style in detail for (Image 1) using the “Free Descr” option at the top of this WF. ## We provides optional feature QWenVL for this purpose.

Author notes from QwenImageEdit2511_Factory2.json · selected release

## If you know you generated (image 1) using SDXL, we recommend enabling this feature. ## Furthermore, if you know the checkpoint and LoRA used during generation, we recommend matching the environment (set via the blue node in the right diagram). ## This allows you to expect a style similar to the original (image 1). ## If you don't know these details, test the quality using the official model “sd_xl_base_1.0.safetensors”. --- ## Configure the red nodes in the right diagram (prompt, KSampler parameters) according to your objectives. ## This example resulted from asking ChatGPT about my specific requirements (1 and 2 below) and iterating until it worked well. ## 1. Qwen's output tends to make skin look overly smooth, so I want to make it closer to the original SDXL. ## 2. I want to prevent degradation even when processing small faces with SDXL.

Download files

Artefacts du modèle

2 artefacts

qwenImageEdit2511FactoryVersion2_v20.zip

zip · 27,3 KB · SHA-256 52dcc8b814a7…1eaf · Civitai

Download file

QwenImageEdit2511_Factory2.json

json · 236 KB · SHA-256 618138590c76…f50e

Download file

Extraits de sources

3 extraits
Note3 nov.civitai.com/models/2264596

Source context: 780 downloads · Type Workflows · Base model Qwen

Note2 oct.civitai.com/models/2264596

Source context: 724 downloads · Type Workflows · Base model Qwen

Note19 juil.civitai.com/models/2264596

Source context: 615 downloads · Type Workflows · Base model Qwen