Using a multimodal training set enhances multi-style and multi-scene capabilities, increasing diversity and strengthening style. The training set does not include human, so its impact on character face is relatively small.
Recommended settings: heunpp2 + linear_quadratic, cfg: 1.5, step: 20, 0.8-1.2
Due to the difficulty of Z-Image training, the step count and CFG need to be increased. However, using the workbench reduces credits significantly; my commonly used setting of 1056x1584 only requires 1.15 credits. I will update this when a better training method is available in future Base versions. Have fun!