I created this because I was curious to see how far training is possible with an insufficient dataset.It can be used with just "scanny, female robot," but if the facial details are not coming out well, please add something like "honeycomb-shape face."Since the retro art style was not completely removed, I think it would be better to use it in conjunction with the other style Lora in the future.

Files with SHA256: 2b2d40d3b5e284cede8b6ecd2c25b816d25225c1c410d8bdc22fef38381c2a5f