<h2>Turbo + Upscale for Qwen Image 2.1</h2>
<div data-youtube-video><iframe width="640" height="480" allowfullscreen="true" autoplay="false" disablekbcontrols="false" enableiframeapi="false" endtime="0" ivloadpolicy="0" loop="false" modestbranding="false" origin playlist src="https://www.youtube.com/embed/MhbXxiCyYQk" start="0"></iframe></div>
<p>This free ComfyUI workflow draws a picture from your prompt in 4 steps with Viggle's turbo LoRA, at 832 x 1248. A detail pass then makes it 1664 x 2496 and draws 4 more steps over it at denoise 0.35. Hair, knit and skin get real texture, and the face stays the same.</p>
<p>The "upscale on / off" switch skips the detail pass when you only want a quick look.</p>
<h3>What you need</h3>
<p>No custom nodes. Tested on ComfyUI 0.37.4. If a node shows red, update ComfyUI: older builds do not have the Qwen Image 2.1 nodes.</p>
<ul>
  <li><code>qwen_image_2.1_int8_convrot.safetensors</code> (7.3 GB) in <code>models/diffusion_models</code></li>
  <li><code>qwen3vl_8b_int8_convrot.safetensors</code> (9.4 GB) in <code>models/text_encoders</code></li>
  <li><code>qwen_image_2.1_vae_bf16.safetensors</code> (676 MB) in <code>models/vae</code></li>
  <li><code>Qwen-Image-2.1-viggle-turbo-4step-lora-r64.safetensors</code> (340 MB) in <code>models/loras</code></li>
</ul>
<p>When you open the workflow, ComfyUI offers to download all four into the right folders.</p>
<h3>First picture</h3>
<ol>
  <li>Drag the workflow onto the ComfyUI canvas.</li>
  <li>Write what you want in the green "Your prompt" box. Full sentences work best: who, what they wear, where, the light, the lens.</li>
  <li>Press Run. The picture is saved in <code>output/StableYogi</code>.</li>
</ol>
<h3>Settings</h3>
<p>4 steps, CFG 1, euler, beta. We rendered the same prompts and seeds on each setting and picked from the pictures. On the simple scheduler a faint second copy of the person showed up on 4 of 11 pictures; on beta, 0 of 11, at the same speed.</p>
<h3>Memory</h3>
<p>We ran it with the graphics memory held to 7 GB and the detail pass on, and it finished. The model files are bigger than an 8 GB card, so plan for 32 GB of system RAM. Out of memory? Set "upscale on / off" to false.</p>
<h3>Licence</h3>
<p>Built with Qwen. Qwen Image 2.1 and the Viggle turbo LoRA are for non-commercial use only. Qwen is licensed under the Qwen RESEARCH LICENSE AGREEMENT, Copyright (c) 2026 Hangzhou Tongyi Laboratory Technology Co., Ltd.</p>
<p>More free workflows: <a target="_blank" rel="ugc" href="https://forgebun.com/go/qwen21wf">Community Resources on my portal</a>.</p>

Turbo + Upscale for Qwen Image 2.1 | ComfyUI workflow, 4 steps

<p>Version 1.0 of the free W11 Turbo + Upscale workflow.</p>
<ul>
  <li>4 steps, CFG 1, euler, beta</li>
  <li>832 x 1248 first picture, 1664 x 2496 after the detail pass</li>
  <li>Upscale on / off switch</li>
  <li>Finished with the graphics memory held to 7 GB</li>
  <li>No custom nodes, tested on ComfyUI 0.37.4</li>
</ul>
<p>Built with Qwen. Non-commercial use only.</p>

Files with SHA256: a274c374847c8b7e6dbddc94fdea469bbc29d582eae969675cff0d0af6d097a0