Unless otherwise stated, all my Anima LoRAs are trained on Base v1.0, so I don't think it should be a compatibility issue. I also tested it at 30 steps without Turbo, and the details of Remi Unoya, Shito Suzukiri (Madana), and Shinozaki Mesana (Madana) were clearly more pronounced compared to the generations without the LoRA.

( I confirmed that the LoRA brings out more detail in each character. However, I haven't checked the accuracy of specific features, such as eye color and hairstyle, in the example generations, so there may be some inaccuracies. )

Could you double-check your sampler and other generation settings?