
This LoRA is effective in turning almost any SD1.5 model into anime style, including realistic models.
It was trained on old anime screenshots, mostly Sailor Moon since it is fairly representative of that era of anime. Apart from Sailor Moon, the training images included Rayearth, Card Captor Sakura, Evangelion and some randoms I could find around the place of sufficient quality.
I realise that almost any Japanese popular style cartoon art is referred to as anime. This ain't that 'anime' generally. The style is more like cel-shaded anime from the late '80s early '90s rather than the detailed vibrant so-called anime you might expect of a doujin cover or artbook.
The style tends to be pastel and dull (even with a good VAE) so you might want to add extra saturation during or post-process.
描述:
训练词语: smanime style
名称: smanime-style-sd15-000008.safetensors
大小 (KB): 4737
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success