bssrdf

Wow! The workaround really works. Thanks.

> @bssrdf could you run [`make style && make quality`](https://github.com/huggingface/diffusers/blob/main/CONTRIBUTING.md#how-to-open-a-pr:~:text=%F0%9F%A7%A8%20Diffusers%20relies%20on,%24%20make%20quality)? Done!

> It still doesn't pass :/ I don't know what's going on. Any way, I rerun ```make style``` and ```make quality``` and here are the outputs. Nothing stands out. ```...

I downgraded ruff and reran the "make style && make quality". Let's give it another push. Yeah, ruff is now on 0.3.5. diffusers uses 0.1.5 which is pretty old.

@standardAI, thanks for the reproducing effort. I just pulled the latest ```diffusers``` main and got the same error. Yes, I did run all ```train_xxx.py``` in the same env and all...

> From the error it looks like the underlying UNet isn't a compatible one. For this script to work properly, we need to have the UNet in a compatible structure....

> @standardAI, thanks for the reproducing effort. I just pulled the latest ```diffusers``` main and got the same error. > Yes, I did run all ```train_xxx.py``` in the same env...

> I see. Would it be possible for you to do the following? > > * Load the UNet from your checkpoint. > * Run a single forward pass on...

Sorry, I mis-read. Yes, [train_dreambooth.py](https://github.com/huggingface/diffusers/blob/37e9d695af0ed1692b8c82fd2af32cb4c0854f81/examples/dreambooth/train_dreambooth.py#L1443) has 1443 lines. Now the log looks like ``` 04/10/2024 07:48:15 - INFO - __main__ - Total optimization steps = 800 Steps: 0%| | 0/800...

Now I see a problem https://github.com/huggingface/diffusers/blob/37e9d695af0ed1692b8c82fd2af32cb4c0854f81/examples/dreambooth/train_dreambooth.py#L1228-L1233 When I run ```train_dreambooth.py```, ```if vae is not None: ``` is False such that ```model_input``` becomes ```pixel_values```. This can not be right. However, when...