Three few-step LoRAs for Qwen Image 2.1 ship within a week of release
linoy_tsaban · x · 2026-09-25
Less than a week after Qwen Image 2.1's release, three few-step LoRAs have shipped. Viggle's turbo LoRA uses DMD distillation to cut the 40-step base to 6 passes for generation and instruction editing (1-3 reference images), roughly 5x faster with near-parity quality — only dense small text lags (8 steps narrows it). Ships as rank-256 (1.3GB) and rank-128 (680MB) adapters with ComfyUI nodes, ready-made workflows, and a 32-example seed-matched comparison Space. Pruna also released 5- and 8-step variants.
More from Multimodal
- Gemini's Call for Me dials phone menus for you, live on Pixel 11 — emmanuelvivier · 2026-09-25
- Gemini 3.8 Live gets real-time video avatar with 97 languages — emmanuelvivier · 2026-09-25
- Krea's video agent turns one prompt into a complete narrative video, no editing — azed_ai · 2026-09-25
- Testing the new BSF face-swap LoRA on Qwen Image 2.1, params included — Dunc4n1d4h0 · 2026-09-25
- Claude Opus draws shockingly realistic image with 45,000 strokes via agent sketchbook — RileyRalmuto · 2026-09-25
- Claude Opus built the Blender layout, Seedance rendered this cinematic short — Gigibossu · 2026-09-25