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

Merged Qwen models with my favourites loras at precise weighing to produce the usual Jib Mix quality.

Files with SHA256: b7e9af07e3c506083b386de344224680fb42a876c4681ff12c81651b40fe66a0