Hi,
I’m impressed with SplitFlow’s inversion-free text-to-image editing. Noticed the paper already demonstrates strong performance with SD3.5 (outperforming baselines in Table 1), but currently only SD3 is listed as a supported model.
Could you add official configuration support for SD3.5—e.g., a dedicated SD3.5_SplitFlow.yaml config file? Thanks!