note from the creator : Based on the previous version of JRU, specifically the model trained with (0.25千界婆 + 0.75JRU21), through targeted training, the performance of lighting and shadows has been greatly optimized. If you feel the lighting and shadows are not good enough, you can also add the word "light" at the beginning of all your prompts. You can still use the artist quality terms from Danbooru to change the art style. In addition, based on my personal test results, the training outcome using Illustrious-XL0.1 as the base model shows performance on Illustrious-XL1.0, whether it's Lora, Lycoris, or something else. Therefore, I might consider directly transferring the training success of this model to Illustrious-XL1.0 in the future. The fusion of 0.25 weight 千界婆娑 and 0.75 weight JRU2.1 has lost many of its characteristics. I originally didn't want to upload it, but my friends all said the generated images looked good, so I decided to share them. You can use the artist tags from Danbooru. I didn’t specifically train it; it's likely that both fused models are based on Illustrious-xl, which is why it doesn't use artist-specific prompts. If you're not using artist prompts, you can directly copy the following prompt strings. If I find time, I might improve the introduction of this update (mainly to showcase the recommended artist prompt strings and their effects, but you can also write your own; after fusion, the usage is essentially the same as Illustrious-xl).