LORA trained on 339 images from yande.re, tagged using WD 1.4 Tagger extension on AUTOMATIC1111's SD WebUI.

• 95%+ NSFW

• Focus on recreating the pink-ish skin tone and softness

As the artist mainly draws ????, increasing the LORA weight will affect clothing somewhat. It also increases the color and the softness of the skin and nipples/vagina.

Tags such as nipples/vagina/etcs will react very aggressive.

The recommened weight range is 0.5~1.0.

3 Versions provided different degrees of strength (Higher Epochs = Stronger), please test with all of them.

Trained and Tested with AOM2-hard + orange VAE, worked quite well with other LORAs.

Images generated with the configurations:

• xformers DISABLED

• Quantization K Samplers ENABLED

• CLIP SKIP 2

• ENSD 31337

• CFG 7.5

• DPM++ SDE Karras 40 Steps

• 512x768

• The Negative Prompts are the same for all images

• The Positive Prompts are the style + trigger words for LORA/situation.

• Other LORAs can be found on civitai by searching the character name

• LORA enabled through Additional Network Extension

• Upscaled with SWIN2SR