tinytensor/Qwen-Image Edit — Change One Thing, Keep the Rest
Instruction-based editing: you say what changes, everything else stays. No mask, no inpainting box, no redrawing the whole frame. The samples show the three edits that work most reliably: a pair of white fox ears and a mask turned into black cat ears with the red outfit untouched, a navy coat recoloured to deep red, and a daytime window replaced with a snowy city night. HOW TO PHRASE IT: name the object and the change, then state what must not move. One change per pass — two unrelated changes in one sentence usually produces one of them, or a blend of both. Anchor the rest of the frame explici
| Author / organization | tinytensor |
|---|---|
| Downloads | 32 (as reported by Civitai) |
| Likes | 0 |
| Task | text-to-image |
| Created | 1 November 2026 |
| Last updated at source | — |
More from tinytensor
- tinytensor/Qwen-Image — Text Inside the ImageModel
- tinytensor/Qwen-Image 2.1 4-Step — 8GB VRAM (1024² in 20s)Model
- tinytensor/4x Upscale — Real-ESRGAN, runs on 8GBModel
- tinytensor/MOSS 2.5D Anime on 8GB VRAM — 1024x1536 in 62s (Apache 2.0)Model
- tinytensor/MOSS Image — Simple Text2Image (8GB VRAM)Model
- tinytensor/Flat Print Style — Taisho Woodblock on Qwen-ImageModel
Silytics shows public metadata from Civitai and links to the original. Information may be incomplete or out of date. Last refreshed 1 November 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.