Jibs Qwen workflow uses a custom-trained Wan VAE to remove the hash/grid lines visible in Qwen outputs. <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:14px">Here is the GitHub for nodes: <a target="_blank" rel="ugc" href="https://github.com/spacepxl/ComfyUI-VAE-Utils?tab=readme-ov-file">https://github.com/spacepxl/ComfyUI-VAE-Utils?tab=readme-ov-file</a> <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:14px">and the Custom Wan/Qwen VAE model: <a target="_blank" rel="ugc" href="https://huggingface.co/spacepxl/Wan2.1-VAE-upscale2x/blob/main/Wan2.1_VAE_upscale2x_imageonly_real_v1.safetensors">https://huggingface.co/spacepxl/Wan2.1-VAE-upscale2x/blob/main/Wan2.1_VAE_upscale2x_imageonly_real_v1.safetensors</a><hr /><span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">V6 a Qwen 2512 Version of Jib Mix Qwen.<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">+ Has better fine details than previous versions.<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">+ Less of a same face issue.<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">+ Less nosie artifacts.<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">- Not quite as good at NSFW as V5. (You can use my NSFW lora to bring it back: <a target="_blank" rel="ugc" href="https://civarchive.com/models/1943554/jibs-qwen-nudity-fixer-lora"><span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">https://civarchive.com/models/1943554/jibs-qwen-nudity-fixer-lora</a><span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">) <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">- Not quite as fine good details as Base Qwen 2512<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">More European faces and fewer Asian by default. The Pruned Model nf4 (13.91 GB) is actually a <a rel="ugc" href="https://civarchive.com/api/download/models/2753478?type=Model&amp;format=GGUF&amp;size=pruned&amp;fp=nf4">Q5_0.GGUF</a><hr />V5 -<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">More realistic, less plastic skin by default, with more imperfections. <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">My Skin <a target="_blank" rel="ugc" href="https://civarchive.com/models/2136360/jibs-qwen-skin-detailerimperfections-lora">detail/imperfections lora</a><span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px"> can be used to add more or even help remove them if used at a negative weight.<span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">The <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/2409636?type=Model&amp;format=GGUF&amp;size=pruned&amp;fp=nf4"><span style="background-color:rgb(26, 27, 30);color:rgb(76, 110, 245);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Version 5 Q5. GGUF (marked nf4)</a> <hr />V4 - Much <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">more natural pretty faces (Much better at Asian faces), less noise, cleaner (slightly less photographic look by default but adding LORAs can make that stronger again) NEW: The <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/2315751?type=Model&amp;format=SafeTensor&amp;size=pruned&amp;fp=fp8"><span style="background-color:rgb(26, 27, 30);color:rgb(34, 139, 230);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:14px">fp8_e5m2</a><span style="background-color:rgb(26, 27, 30);color:rgb(76, 110, 245);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:14px"> <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:14px">model is twice as fast as the Q5.GGUF on my 3090. There is a <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/2315751?type=Model&amp;format=GGUF&amp;size=pruned&amp;fp=nf4">Q5 .GGUF (marked nf4)</a> <span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Tune for Clownshark dpmpp_3s/Bong_tangent sampler this time instead of Euler_ancestral/Linier_ Quadratic. <hr />V3- Important I have uploaded 2 different versions of this model: <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/2268551?type=Model&amp;format=SafeTensor&amp;size=full&amp;fp=fp16">High Noise version (marked fp16)</a> that I think is better at lower steps and single stage workflows (but can sometime show grid/scan lines). <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/2268551?type=Model&amp;format=SafeTensor&amp;size=full&amp;fp=fp16">Lower noise version (marked fp32)</a> That makes cleaner/less noisey images on the first gen but is better suited to 2 stage Hi-res fix workflow (this is my preferred method) There is a <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/2268551?type=Model&amp;format=GGUF&amp;size=pruned&amp;fp=nf4">Q5 .GGUF (13.91 GB)</a> that is the Higher Noise version The <a target="_blank" rel="ugc" href="llama.cpp\build\bin\Debug\llama-quantize.exe C:\ComfyUI_For_Wan\ComfyUI-Easy-Install\ComfyUI-Easy-Install\ComfyUI-Easy-Install\ComfyUI\output\diffusion_models\Jib_Mix_Qwen-Image_V3_A_00001_-BF16.gguf C:\ComfyUI_For_Wan\ComfyUI-Easy-Install\ComfyUI-Easy-Install\ComfyUI-Easy-Install\ComfyUI\output\diffusion_models\Jib_Mix_Qwen-Image_V3_Q5_K.gguf Q5_K">Q6 .GGUF (15.6 GB)</a> is a lower noise realistic skin version.Small <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/2268551?type=Model&amp;format=GGUF&amp;size=pruned&amp;fp=bf16">Q5 .GGUF (13.91)</a> is a lower noise realistic skin version.<hr />V2 - I tried to fix the big bobble heads from the previous versions. It is better but they can still be a bit big sometimes. The V2 fp8 model listed is now actually a <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/2226001?type=Model&amp;format=GGUF&amp;size=pruned&amp;fp=fp8">Q8 .GGUF </a>(Thanks to export_tank_harmful for the conversion when mine was broken) <hr />V1 - Adds a more photorealistic/amateur look to Qwen. The fp8 model listed is now actually a <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/2192261?type=Model&amp;format=GGUF&amp;size=pruned&amp;fp=fp8">Q6_k .GGUF</a>and has fewer grid lines than the fp8 (But I still recommend the fp16 with ram offloading, for better quality)

Jib Mix Qwen

jibMixQwen_v30.safetensors

jibMixQwen_v30.gguf

v3.0

Jibs Turbo 2512 Workflow

Jibs Qwen 2512 Workflow

v6.0

Jibs Qwen Workflow

v5.0

v4.0

v2.0

v1.0 - Realistic Qwen

Files with SHA256: 2432d89bc694efeee0feabed18f2d7adff422accaff13ce5bf9b984867c5cf7c