Mai Shiranui and Kasumi DoA outfit. All images are generated with A1111 on PC. Trained with Kohya with 80 realistic pics and tested with my Illustrious model available on TA/HUB.KlingAi bonus code : <a target="_blank" rel="ugc" href="https://pro.klingai.com/h5-app/invitation?code=7B4J8NUA5J4X">https://pro.klingai.com/h5-app/invitation?code=7B4J8NUA5J4X</a><span style="background-color:rgba(34, 139, 230, 0.15);color:rgb(254, 254, 254);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">Unfortunately, I am forced to stop sharing my models on this site. The reasons are mainly four. The first is the lack of support from users who download the models only after the "early access." Due to the costs of training and generation on Civitai, without support I cannot continue. Another reason is the low number of votes, many download without ever rating the model. The most serious reason is that many have authorized themselves to publish my models on their sites or in other groups like Pixai, Tensor, etc., and some of them sell them on Patreon or other sites like DeviantArt. At this point, I am stopping publishing on Civitai and continuing on Tensor with the download blocked. You can follow me and test my new releases on TensorHub.. Thank you. TensorHub : <a target="_blank" rel="ugc" href="https://tensorhub.art/it-IT/u/726760736508024852">https://tensorhub.art/it-IT/u/726760736508024852</a>This model does not use any real-life or celebrity content. Please do not use this model for any real-life, celebrity, or pornographic content.Illustrious Weight 0.8 CFG 5 on realistic checkpoints.If you like the character, always leave a rating and a picture, and if you give some buzz you will allow me to train other LORAs or MODELS. Downloading only is not useful. For model visibility, remember to rate the images in my gallery. Thank you.IMPORTANT: All my models work only if the trigger words and the recommended negative and positive prompts and embeddings are used. If you have a different result from the images, it is because you have not used the recommended settings. Too many loras and/or embeddings can ruin the model's result. To have images like the ones I made, use ONLY the recommended settings. If you use the indicated settings (below), you should be able to create images like the ones shown in the photos/videos. The delicate images with an ethereal gradient are intentional and included in the lora. With anime models, it could give completely different results. In the images, I used my semi-realistic model available on TA..Example positive prompt:(realistic), masterpiece, 4k, detailed, photo of an asian woman kasumikun, brown hair, brown eyes, ponytail, yellow ribbon, hair ribbon, choker, blue kimono, japanese clothes, short sleeves, (pelvic curtain), cleavage, blue arm guards, white thighhighs, shin guards, blunt bangs, large breasts, 1girl, happy, narrow waist, big thighs, large hips,  pose in a garden...For back view, add : from behind, looking back(realistic), masterpiece, 4k, detailed, photo of a woman shiranuimai, brown hair, brown eyes, long high ponytail, hair tie, makeup, red japanese clothes, sleeveless, bare shoulders, large breasts, cleavage, 1girl, solo, hand on hip, pelvic curtain, ninja, shiranui mai, happy, eyelashes, slender body, big thighs, large hips , pose in a garden...<h3 id="settings-for-illustrious-version:">Settings For Illustrious version:</h3>IMPORTANT, PLEASE READ : If you get different results, the prompt was probably not followed and an unsuitable checkpoint was used. All my LoRAs have been trained on realistic models. The videos I post represent what you can achieve; even though the video is animated, it reflects the quality and what the LoRA can do. The first frame of the video is the image I created using the LoRA and then animated, using the parameters I recommended. To achieve a realistic result, a <a target="_blank" rel="ugc" href="https://tensor.art/it-IT/models/869285461682562510/Illustrious-SDXL-Realistic-v1">realistic</a> ILLustrious non-merged checkpoint should be used. If you use merged models with anime models, the result may be very different from mine. Do not use corrective loras in addition to the main lora because they often only ruin the result. If possible, do not exceed 150 tokens in prompts. Too many words in prompts do not improve the result. The parameters I use are the following :Steps: 35, Sampler: Euler a, Schedule type: Automatic, CFG scale: 4.5, Size: 512x768, Denoising strength: 0.35, ADetailer model: <a target="_blank" rel="ugc" href="https://huggingface.co/Bingsu/adetailer/blob/main/face_yolov9c.pt">face_</a><a target="_blank" rel="ugc" href="http://yolov9c.pt">yolov9c.pt</a>, ADetailer confidence: 0.5, ADetailer dilate erode: 4, ADetailer mask blur: 8, ADetailer denoising strength: 0.25, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer use separate CFG scale: True, ADetailer CFG scale: 4.5For Adetailer face, you can choose the better: face_yolov10x_<a target="_blank" rel="ugc" href="http://face.pt">face.pt</a>Reposting on other sites is prohibited.It is strictly prohibited to distribute outside civitai this model or hosting services in third-party image generation services, or sell my images on other site. You can use this model for personal use on Civitai or A1111 on your pc. Not for sell or repost on pixai or other sites.If you appreciate the lora, please consider leaving us a review and sharing the images generated using this model. Your feedback is valuable to both us and the AI community. Thanks.

Files with SHA256: 00981e63d40d624f263ae7f39f420bddb4c2a04de135a742536076cda1b1e6f6