-
I HIGHLY RECOMMEND! Use LoRA with a strength of 0.6 and lower. When setting the strength higher, distortions begin to appear, which are clearly visible when generating the same image without using LoRA.
-
The trigger word is "drawn in nlwx style" or "nlwx style". But I still recommend the first one.
LoRA trained on the works of the artist "Alexander_Dinh". I don't know if it was a good idea to take his works for my first LoRA, but after training two LoRAs I realized that his works are not very suitable for training (or my knowledge is not enough, which is also quite possible), since he uses mixed media, there can be both 3D objects and just photos or screenshots from games in the background, and very poorly drawn elements, especially hair and clothes.
描述:
The training materials and quantity have been greatly improved, but I still made a few mistakes :P
训练词语: drawn in nlwx style ,nlwx style
名称: Alexander_Dinh_Style_v095_(QuantDweller).safetensors
大小 (KB): 90905
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success