qwen-image-edit

1# Qwen Image Edit Workflows
2 
3## Overview
4 
5Qwen Image Edit uses a vision-language model (Qwen2.5-VL) to edit images based on natural language instructions. The model "sees" the source image through CLIP conditioning and generates an edited version.
6 
7## Models
8 
9### Required Components
10 
11| Component | Node | Model Name | Notes |
12|-----------|------|------------|-------|
13| **UNET** | `UNETLoader` | `qwen_image_edit_2511_bf16.safetensors` | Official 2511 edit model (bf16) |
14| **CLIP** | `CLIPLoader` (type=`qwen_image`) | `qwen_2.5_vl_7b_fp8_scaled.safetensors` | Shared across all Qwen models |
15| **VAE** | `VAELoader` | `qwen_image_vae.safetensors` | Qwen-specific VAE |
16 
17### Alternative UNET Models
18 
19| Model | Path | Focus |
20|-------|------|-------|
21| `qwenImageEditRemix_v10` | `qwenImageEditRemix_v10.safetensors` | Community remix, general editing |
22| `qwenUltimateRealism_v11` | `Qwen/imageized/qwenUltimateRealism_v11.safetensors` | Product photography, hyper-realistic |
23| `copaxTimeless` | `Qwen/realistic/copaxTimeless_qwenUltraRealistic.safetensors` | Ultra-realistic portraits |
24| `qwnImageEdit_v16Bf16` | `Qwen/abliterated/qwnImageEdit_v16Bf16.safetensors` | Abliterated (uncensored) |
25 
26## Conditioning Nodes
27 
28### TextEncodeQwenImageEditPlusAdvance_lrzjason (Recommended)
29 
30From the `qweneditutils` custom node pack. The **Advanced** variant is preferred because it:
31- Outputs a **LATENT directly** (no need for separate EmptyLatentImage)
32…