A LoRA (maybe slightly over)trained on images mostly from Viper V16, as well as Viper GT1 and GTB via the NAI model. It's a small dataset comprised of lowres images, but the results are pretty good considering. Can also recreate the characters and outfit to an extent—some better than others.
The waitress outfit: waitress uniform, underbust, skirt, ribbon
Akira: akira, red hair, very long hair, red eyes, (maybe also 'high ponytail')
Karin: karin, blonde hair, twin braids, blue eyes, glasses
Saki: saki, blue hair, blue eyes — then either 'long hair, ponytail' OR 'short hair'
Makoto: makoto, dark-skinned female, black hair, blue eyes, headband
Narsha: narsha, orange hair, short hair, red eyes, forehead jewel, antennae, hairpods
Saki and Narsha are less consistent than the others; Saki because she has multiple hair lengths/styles in the dataset, and Narsha because she's Narsha, so good luck if you're trying to prompt them!
You can also find me on pixai.
https://pixai.art/@thereal-ainonymous
描述:
训练词语:
名称: viper-20.safetensors
大小 (KB): 36993
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
