New face swap LoRA for Qwen Image 2.1 impresses with identity transfer, though style matching lags
linoy_tsaban · x · 2026-09-28
A hands-on test of @Alissonerdx's new face swap LoRA for Qwen Image 2.1 shows impressive identity transfer, lighting and pose reproduction.
Caveats: style matching to the reference/target images could be better, and it doesn't preserve the target's hairstyle — which the tester jokes borders on a philosophical discussion of visual identity. The LoRA is available to try online.
More from Multimodal
- Ad Buyer Ditches Image Models, Uses Seedance 2.5 for Realistic AI UGC Characters — churchkey · 2026-09-28
- One-word Midjourney prompt: 'Pulchritudinous' at --ar 5:4 --v 8.2 — tisch_eins · 2026-09-28
- Prompt Template Makes Qwen Image 2.1 Design Like Closed-Source Models — wjc_5 · 2026-09-28
- A ChatGPT prompt turns GPT Image into an editorial fashion photographer — umesh_ai · 2026-09-28
- Synthesia ships Express-3, its best avatar model, free for users on all plans — synthesiaIO · 2026-09-28
- Alibaba open-sources Qwen-Image-2.1: 7B image model with native RGBA output, claimed to beat closed rivals — thione · 2026-09-28