Question about the training.

[congwei1230]

Thanks for the great work! Regarding the segmentation_mask_brushnet_ckpt_sdxl_v0 and random_mask_brushnet_ckpt_sdxl_v0 provided on Google Drive, Could you please provide us with information about the number of training steps completed for each of these checkpoints? I assume that these models are trained with a batch size of 1 and a gradient accumulation of 4 on 8 GPUs. Thank you! @juxuan27