<h1 id="important-to-read:"><span style="color:rgb(250, 82, 82)">Important to Read:</span></h1><p></p><h2 id="qwen-image-2512:"><span style="color:rgb(190, 75, 219)">Qwen-Image-2512:</span></h2><p></p><h3 id="check-on-&quot;about-this-model&quot;-to-find-the-links-of-the-models-recommended-to-achieve-this-quality-unfortunately-any-type-of-quantization-downgrades-the-quality-and-worsen-the-prompt-adherence-i-recommend-using-my-workflow-or-adding-the-upscaling-step-to-your-own-workflow.-all-images-were-upscaled-and-i've-used-the-seedvariance-node.-on-an-rtx-4060-ti-it-takes-around-2m15s-every-image-including-the-upscaling.">Check on "about this model" to find the links of the models recommended to achieve this quality, unfortunately any type of quantization downgrades the quality and worsen the prompt adherence, I recommend using my workflow, or adding the upscaling step to your own workflow. All images were upscaled and I've used the seedvariance node. On an RTX 4060 ti it takes around 2m15s every image, including the upscaling.</h3><h3 id="recommended-weights:-(0.80-1.00)">Recommended weights: (0.80 - 1.00)</h3><h3 id="trigger-words-&quot;erect-penis-exposed&quot;-&quot;bare-ass&quot;.">Trigger words "erect penis exposed", "bare ass".</h3><h3 id="a-recommendation:">A recommendation:</h3><h3 id="it-would-be-better-to-explore-and-give-different-instructions-to-the-model-to-represent-directions-further-from-my-suggestions.">It would be better to explore and give different instructions to the model to represent directions, further from my suggestions.</h3><h3 id="&quot;erect-penis-exposed-(orientation_back)&quot;-to-show-the-penis-from-the-back-perspective-it-may-not-work-perfectly-and-sometimes-it-works-better-only-using-&quot;bare-ass&quot;.-there-are-also-(orientation_frontsideupdown)-but-it-doesn't-seem-to-be-followed-at-all-but-you-are-free-to-test-it.">"erect penis exposed (orientation_back)" to show the penis from the back perspective, it may not work perfectly, and sometimes it works better only using "bare ass". There are also (orientation_front/side/up/down) but it doesn't seem to be followed at all, but you are free to test it.</h3><h3 id="it-took-around-21-hours-of-training-(onetrainer)-512-resolution-152-images.">It took around 21 hours of training (Onetrainer), 512 resolution, 152 images.</h3><h3 id="it-can-cause-deformation-and-unwanted-results.-and-i-know-the-two-2d-outputs-aren't-ideal-but-it-has-a-little-flexibility-for-2.5d-styles-(as-much-as-the-base-does).-and-it-does-not-perform-sex-or-sucking-it-can-do-sucking-sometimes-but-mostly-can't.">It can cause deformation and unwanted results. And I know the two 2D outputs aren't ideal, but it has a little flexibility for 2.5D styles (as much as the base does). And it does not perform sex or sucking, it can do sucking sometimes but mostly can't.</h3><p></p><h2 id="zimage-turbo:"><span style="color:rgb(130, 201, 30)">Zimage-Turbo:</span></h2><p></p><p></p><h2 id="to-get-the-best-results-you-need-to-use-this-lora-in-order-for-it-to-work-properly.">To get the best results, you need to use this <a target="_blank" rel="ugc" href="https://huggingface.co/thwri/the-bin/blob/main/z-image/z-image_turbo_distillpatch_comfyui.safetensors">LoRA</a> in order for it to work properly.</h2><p></p><h3 id="recommended-strength-(1.00-1.20)">Recommended strength (1.00 - 1.20)</h3><p></p><h2 id="for-optimal-results-use:-&quot;erect-circumcised-penis-exposed&quot;-but-&quot;erect-penis-exposed&quot;-should-work.-i-recommend-to-use-the-workflow-on-the-images-to-achieve-high-quality-outputs."><span style="color:rgb(134, 142, 150)">For optimal results, use: </span><span style="color:rgb(34, 139, 230)">"erect circumcised penis exposed"</span><span style="color:rgb(134, 142, 150)"> but </span><span style="color:rgb(34, 139, 230)">"erect penis exposed"</span><span style="color:rgb(134, 142, 150)"> should work. I recommend to use the workflow on the images to achieve high quality outputs.</span></h2><p></p><h3 id="the-samples-were-generated-with-the-&quot;qwen3-4b&quot;-text-econder-but-for-optimal-results-i-recommend-&quot;qwen3-instruct-2507-4b&quot;-since-it-was-trained-with-it.">The samples were generated with the "Qwen3 4b" text econder but for optimal results I recommend "Qwen3 instruct 2507 4b" since it was trained with it.</h3><p></p><h3 id="it-took-less-than-6-hours-of-training-(onetrainer)-512-resolution-94-images.">It took less than 6 hours of training (Onetrainer), 512 resolution, 94 images.</h3><p></p><h3 id="as-always-remember-that-the-model-can-produce-some-unpleasant-results-struggle-with-some-prompts-and-in-weird-cases-cause-deformation.">As always, remember that the model can produce some unpleasant results, struggle with some prompts, and in weird cases, cause deformation.</h3><p></p><p></p>

The Cock Z

<p>I recommend using qwen_image_2512_fp8_e4m3fn_scaled_comfyui_4steps_v1.0, But using 8 steps instead of 4. Here is the link:</p><p><a target="_blank" rel="ugc" href="https://huggingface.co/lightx2v/Qwen-Image-2512-Lightning/blob/main/qwen_image_2512_fp8_e4m3fn_scaled_comfyui_4steps_v1.0.safetensors">https://huggingface.co/lightx2v/Qwen-Image-2512-Lightning/blob/main/qwen_image_2512_fp8_e4m3fn_scaled_comfyui_4steps_v1.0.safetensors</a></p><p>The text encoder I've used to train this LoRA was R1 Onevision 7B, you can download any of the GGUF version files here:</p><p><a target="_blank" rel="ugc" href="https://huggingface.co/mradermacher/R1-Onevision-7B-GGUF/tree/main">https://huggingface.co/mradermacher/R1-Onevision-7B-GGUF/tree/main</a></p><p>The VAE (optional, I'm note sure if it changes anything):</p><p><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/blob/main/split_files/vae/wan_2.1_vae.safetensors">https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/blob/main/split_files/vae/wan_2.1_vae.safetensors</a></p><p>The model for upscaling:</p><p><a target="_blank" rel="ugc" href="https://openmodeldb.info/models/1x-SwatKatsLite">https://openmodeldb.info/models/1x-SwatKatsLite</a></p><p>Credits for the upscaling nodes I've used:</p><p><a target="_blank" rel="ugc" href="https://openart.ai/workflows/caiman_thirsty_60/flux-dedistilled-ultra-realistic-t2i-dual-upscale-facedetailer/1Hhrjm54PjM632btIY3A">https://openart.ai/workflows/caiman_thirsty_60/flux-dedistilled-ultra-realistic-t2i-dual-upscale-facedetailer/1Hhrjm54PjM632btIY3A</a></p>

Files with SHA256: 0673a02091e9880a7792fd2f77b038b44b3719bf5d311579f25c2ffefcc11598