
This is a style LoRa hatched from an idea I had a couple months ago. Take stuff I like, caption it, train it. No real connective tissue between things, just general stuff I thought was sexy or cool. It's kind of fun, I think. Because I used captions, you might be able to organize the chaos.
There's no rhyme or reason to what I trained on, but you can expect:
- woman and porn because, duh, I'm a pervert. Expect an emphasis on curvy women, women of color, and curvy women of color. This includes some manga and hentai.
- still life food photography, primarily sweets and junk food with some fruit. I am not here to show you kale.
- street photography and urban landscapes
I plan to train future versions incrementally. I thought about doing them monthly but the tagging for this was a major pain. It took me almost a week to caption these. So also expect less aggressive cadence but also less diligent tagging in the future.
描述:
I trained on all 303 images from both training sets. This replaced a version that was up briefly where I just merged the two Loras; retraining on the combined set turned out better,
训练词语:
名称: OttoZone_v2+.safetensors
大小 (KB): 73877
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success