Dataset
#26
by
richardshkim
- opened
I noticed that LAION was used for other SDXL-ControlNet models, but I couldn't find any information about the dataset used for SDXL-Inpainting on the model page. Could you please clarify which dataset was used for training SDXL-Inpainting?
Thank you!
Hi. Unfortunately, I don't know what this was trained on as I'm a newer member on the team, and wasn't at HF when this was made. Maybe someone else from the team can help answer
I am also curious about the dataset.
Also, in the model description: During training, we generate synthetic masks and, in 25% mask everything.
Is the "mask everything" means the mask is a pure white image?