When I tried out a LoRA of an artist called Doyora (one of my favorite artists) I got from Catto's LoRA archive on GitGud, I noticed it wasn't that great with recent models such as AURORA (I highly reccomend you download and use it since it's the best anime model I've ever used in terms of versatility and prompt accuracy), surely because it could be lots of months old, so I decided to train one my own but as a LoCon after following a guide here made by Hollow/Holostrawberry.
Since this artist is known for his (kind of) oil painting style, depending on the model you use, you may not get the exact style of his (more or less) from the images of the dataset I trained it with, like for me with AURORA, I mostly get semi-realistic style by default on my end, let me know if using another model does give you his exact style or not.
描述:
训练词语:
名称: Doyora.safetensors
大小 (KB): 38185
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
