How about sd3 lora trainning?

Thanks for this swif update for sd3 dreambooth trainning, but how about the implementation of sd3 lora?
It seems the latter one is best choice for small-scale datasets.

Review the existing SD3 DreamBooth training implementation mentioned in the issue and determine how SD3 LoRA training would fit alongside it. The issue does not name files, tests, or entry points; completion would require implementing and validating SD3 LoRA training support for small-scale datasets.

Written by the indexing model from the issue text.