JANGSOONMYUN

@mona0809 Thank you for your reply. Ok I will test it.

@danielS91 Thank you. I checked the NYUV2 dataset and there is 'void' in the classes. If there is unseen object detected, is it classified as 'void' ?

@mona0809 I see. Maybe I should train the model with additional dataset and see the result. I appreciate.

> Hello @loboere I tryed with --use_8bit_adam and got bad results as well, but with different params my results were better. > > > accelerate launch dreambooth_inpaint.py ^ > >...

> @JANGSOONMYUN you have to modify the code to support your data, I suggest you take a look at the text-to-image script [here](https://github.com/huggingface/diffusers/blob/main/examples/text_to_image/train_text_to_image.py) Ok thank you!

> I guess you can download the weights file from this drive link, probably the old links are broken. https://drive.google.com/drive/folders/1xu2Pq6mZT5hmFgiYMBT9Zt8h1yO-3SIp Thank you for the weight files. However, with "indoor_ds_new.ckpt" the...

> Hi > try to replace with > > from .lr_scheduler import LR_Scheduler I just copied the lr_scheduler.py file and renamed lr_scheduler_orig.py, and also changed class name to LR_Scheduler. Somehow...