In a Class of Her Own is a 2020 Chinese television series starring Ju Jingyi, Song Weilong, Bi Wenjun and Wang Ruichang. The series premiered on iQiyi with multi-languages subtitles globally on July 23, 2020.<h2 id="also-check-the-sd1.5-model-(768x768)-2itnto54m"><a target="_blank" rel="ugc" href="https://civarchive.com/models/120846/in-a-class-of-her-own-yu-lexuan?modelVersionId=131453">Also check the SD1.5 model (768x768)</a></h2><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/e52c2a1c-e490-411a-afe2-8641d6acca21/width=525/e52c2a1c-e490-411a-afe2-8641d6acca21.jpeg" />This highly experimental model was trained in a small dataset of only 11 high-quality images. A new version with at least 60 images is being refined. The v1.5 model offers more flexibility.<h2 id="usage-tips:-84dcy9dit">Usage tips:</h2>♦The activation tag is just "yulexuan', if you can't get a good likeness of him, "asian male", "wearing a small coronet made of fabric" and "wearing blue hanfu" also work.♦Set clip skip to 1.♦Include "transparent background" and "black background" in the negative prompt. An error while cleaning the PNGs passed these attributes to the AI and it will show up if lora strength is above 0,7.♦Resolution of 1024 x 1024, 1024 x 768 or near is also fine as these were the resolutions used for training.♦LoRa strength can be anywhere from 0.3 to 1. Best results seem to be at 0.6 to 0.8 depending on model.♦It works best on vanilla SDXL and Realistic Vision XL. Good results on OpenDall-e as well.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/79a7da68-4638-47c7-bf17-680c2d950528/width=525/79a7da68-4638-47c7-bf17-680c2d950528.jpeg" />♦Use <a target="_blank" rel="ugc" href="https://civarchive.com/articles/3274/complete-guide-on-how-to-use-adetailer-after-detailer-all-settings-explained">ADetailer</a> instead of face restoration.♦It has been trained using the base SDXL 1.0 model, so it should be compatible with all checkpoints. This is a photorealistic model, and won't work on Pony.This model is not filtered and may produce NSFW.<h2 id="description-rye322cki">Description</h2>This is the SDXL LoRa version of <a target="_blank" rel="ugc" href="https://civarchive.com/models/78305/yu-lexuan">my first checkpoint </a>ever trained, Yu Lexuan:<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/aa73fa02-093c-4811-abdd-699399155a2b/width=525/aa73fa02-093c-4811-abdd-699399155a2b.jpeg" />This Lora has been optimized to reproduce his most recognizable clothing, a nice blue hanfu. It does a much better job than the checkpoint.This SDXL version uses a reduced dataset with just 11 high-quality images scrapped from Weibo. I have close to 100 images with varied quality.

Files with SHA256: 5a8635d07c09c392ee16e596827e4de27e3ead790151f37d8744b0f0fd936406