Viggle Releases 4-Step Turbo LoRA for Qwen Image 2.1
Viggle released a Turbo LoRA for Qwen-Image 2.1 (Qwen-Image-2.1-viggle-turbo) that uses DMD distillation to compress sampling to 4 steps, with roughly 6x speedup and quality that holds up in side-by-side comparisons. Weights are open on Hugging Face with an online demo Space, and the model climbed both the HF trending and Space hot charts.
Confirmed
- Weights are freely downloadable on Hugging Face (Viggle/Qwen-Image-2.1-viggle-turbo), alongside a Gradio demo Space requiring no local setup
- Positioned as a text-to-image / image-editing pipeline on top of Qwen-Image 2.1, focused on fast inference
- 4-step generation with an officially claimed 6x speedup at near-original quality
- Reached both the Hugging Face trending list and the Space trending list
- Reddit user @ExternalQuarter noted the turbo variant appeared on Hugging Face surprisingly soon after the base model's release
Why it matters
- 4-step generation sharply cuts inference cost and wait time, directly useful for fast-iteration design and editing workflows
- Released as a LoRA rather than a standalone model, so users can flexibly stack it onto Qwen-Image 2.1
- It illustrates how quickly the community now ships distillation speedups on top of new base models
2026-09-23 ~ 2026-09-25 · 5 related posts
Primary sources
- Qwen Image 2.1 Turbo quietly appears on Hugging Face — External_Quarter · 2026-09-23
- Qwen-Image-2.1 distilled into 4-step turbo, 6x faster with comparable quality — linoy_tsaban · 2026-09-23
- [source] Viggle releases 4-step Qwen Image 2.1 Turbo LoRA with open weights and demo — fruesome · 2026-09-23
- [source] Viggle's distilled Qwen-Image 2.1 turbo LoRA trends on Hugging Face — Viggle · 2026-09-24
- [source] Viggle's distilled Qwen image-editing turbo model trends on Hugging Face — Viggle · 2026-09-25