I train this lora by anima_baseV10, but after testing, it's work better by using animayume_v05. Maybe you can try other checkpoints to test its the most suitable model....
我用anima_baseV10训练的该模型，但实际上经过测试它在animayume_v05 这个微调底膜上效果更好.你可以试试其他的anima微调底膜说不定会更合适？
Recommend: CFG(Scale): 5, 25~30 steps, and Euler a as the sampler,and Strength:0.7~0.9(Because the dataset this lora will influence hands,0.7 will reduce its effect ) .
lora强度我建议维持在0.7~0.9，过高的强度会影响手的绘画，目前我的测试是0.7最佳，在往上有概率崩手