這是美女網紅並帶有"NSFW"內容的LORA,全部訓練圖片都是裸體,但無法在這展示,與平常一樣寫出私密部位的提示辭即可。對於想要在"Tensor Hub"上使用的朋友非常友善,只需用這網站帳號登入"Tensor Hub" ,再搜尋我的三個帳號即可看到"NSFW"內容。這次訓練的人物針對皮膚部分加強處理,看上去會變得光滑細膩水潤質感。

This is LORA, a beautiful internet celebrity with "NSFW" content. All her training photos are nude, but cannot be displayed here. As usual, you can use the descriptions of her private parts. It's very user-friendly for those who want to access it on "Tensor Hub." Simply log in to "Tensor Hub" with your "Tensor Art" account and search for my three accounts to see the "NSFW" content. The skin of the person in this training session has been enhanced, appearing smooth, delicate, and moisturized.

推薦採樣器（sampler）： Euler a / ddim_uniform (推薦recommend)

這裡特別說明：使用Euler a採樣器會改善膚質,而使用ddim_uniform調度器是因為人物姿勢及表情多變,當然其他調度器也能使用,但使用過就知道效果.

A special note here: using the "Euler a" sampler improves skin texture, while the "ddim_uniform" scheduler is used because character poses and expressions are highly variable. Of course, other schedulers can also be used, but you'll know the effect once you've tried them.

步數（steps）：9-11

觸發辭（Trigger words）：celebrity

權重（Strength）：0.8-1.0

觸發詞可用可不用,人物相似度極高。但不用觸發詞好處是繪圖"NSFW"圖片時,比較不會有衣服干擾,老司機都懂得。

The trigger words are optional, and the characters are extremely similar. However, the advantage of not using trigger words is that when drawing "NSFW" pictures, there will be less interference from clothing, as experienced viewers will understand.