Qwen-Image-Edit 2511 4-Step Fast
__GITHUB_SVG__ GitHub
No images
Upload: Max 2 images  ·  #1 Main / #2 Reference
Output
__DOWNLOAD_SVG__ Save
Quick Prompts
Editing Style / LoRA ? A LoRA is a specialized skill pack for the editor. Pick one that matches your goal, then write a matching prompt.
A LoRA is a specialized skill pack for the editor (style, camera, lighting, upscale, etc.). Pick one that matches your goal, then write a prompt that asks for that kind of change.
Before
Before demo
After
After demo
Demo for Choose a LoRA…
Processing image…
Result will appear here
Edit Instruction
Prompt Packs add variety the base model struggles with in specific areas. Choose a pack in the menu below, or trigger one in the prompt (e.g. [[pack]]).
Prompt Pack ? Optional fixed instructions + reference image for things the base model under-knows. Menu selection overrides prompt tags. If a pack sets force_lora, the LoRA menu switches to that adapter (and resets to None when you clear the pack). Choose None to use tags in the prompt (e.g. [[example]]). Packs load from the prompt_packs/ folder.
Output Dimensions / Ratio ? Sets the output shape. Original keeps the source proportions; Square is 1:1; Wide is 16:9 landscape; Portrait is 9:16 vertical. This only changes the frame ratio, not overall resolution.
Output Image Size ? Controls output resolution. Original uses source pixels (long side max 4000). Ladder: X-Small 512, Small 768, Normal 1024, Large 1280, X-Large 1536 on the long side. Larger images use more GPU memory and take longer.
0
1.0
4
Experimental Space for Qwen-Image-Edit-2511 · LoRAs loaded lazily on first use
No images uploaded
Ready