Noct Q

Qwen Image 2.1, uncensored out of the box, packaged for ComfyUI in two versions: Turbo (5 steps) and Base (full steps). Nudes and explicit scenes render without a LoRA.

What's new in V3

• Turbo rebuilt on Viggle's v0.2.1 distillation, sampled in 6 steps. sharper images and seeds as varied as the full model

• In a blind test on the same 150 prompts and seeds as V2 Turbo, less defect with the same prompt adherence and explicit rendering

• Base is a new model based on three community NSFW LoRAs

Settings V3

• Turbo: 6 steps, euler, no CFG, 832x1248. Larger sizes work but bodies break more often.

• Base: 25–50 steps, euler, simple. V3 Base: cfg 3 with the negative prompt, both set in the workflow.

What's new in V2

• Built on Viggle's v0.2 distillation.

• The leftward shift in edits is gone.

• Turbo: 5 steps, no CFG, about 11 s for 832x1248 on an 8 GB card

• Base: the undistilled model, 25-50 steps, for the most faithful prompt following and for fine-tuning

• The workflows use core ComfyUI nodes only

What it does well

• Full-body nudes in daylight, flash and night scenes

• Explicit scenes, oral most reliably

• Readable text on signs and labels

• Anime and painted illustration

• Plain-caption prompts: who, what she does, where, the light

• Speed: about 12 s for a 1024x1536 image on an 8 GB card with int8 convrot

• Image editing with the same files, in ComfyUI's Qwen Image 2.1 edit template

Which file by Vram

• 8–12 GB: int8 convrot (7.3 GB), the default, almost pixel-identical to bf16

• 6–8 GB: int4 mixed (4.6 GB), the fastest; compositions shift a little

• GGUF Q8_0, Q6_K, Q5_K, Q4_K (7.6 to 4.1 GB)

• fp8 (7.3 GB): plain fp8 e4m3fn

• bf16 and fp16 (14.2 GB): for merging and training; fp16 for GPUs without bf16

Settings

Turbo: 5 steps, euler, no CFG, on the schedule set in the workflow (core ManualSigmas node), 832x1248. Larger sizes work but bodies break more often; the workflow note gives the schedule for each size.

Base: 25–50 steps, cfg 1, euler, simple.

Notes

• The negative prompt does nothing at cfg 1

• Based on the 5-step distillation by Viggle

• Shibari is VALIDATED, may allow obscure position support don't hesitate to try

• Don't use it on real people.

中文：基于 Qwen Image 2.1，开箱即可生成裸体与成人场景，无需 LoRA。两个版本：Turbo（5 步，无 CFG，采用 Viggle v0.2 蒸馏）和 Base（完整模型，25–50 步，cfg 1）。8–12G 显存用 int8 convrot，6–8G 用 int4 mixed。仅限非商业用途。

License

Built with Qwen. Qwen is licensed under the Qwen RESEARCH LICENSE AGREEMENT, Copyright (c) 2026 Hangzhou Tongyi Laboratory Technology Co., Ltd. All Rights Reserved. Free, non-commercial use only. Full licence: https://huggingface.co/Qwen/Qwen-Image-2.1/blob/main/LICENSE. LICENSE and NOTICE, with the list of changes, are in the zip.