Loading into pipeline from single file

Hello! I am new to diffusers and when I try to use the from_single_file() method for the UNet2DConditional model, I was getting errors:

unet = UNet2DConditionModel.from_single_file( "https://huggingface.co/CompVis/stable-diffusion-v1-4/unet/diffusion_pytorch_model.safetensors")
I tried it with the .bin file and also using a local file after downloading.
All of them give the same error.
SingleFileComponentError: Failed to load UNet2DConditionModel. Weights for this component appear to be missing in the checkpoint.

Can someone explain why this is happening and what is expected?

There is currently a highly politicized, perhaps, but technically very shabby problem with from_single_file.
It might work if you download the config from the appropriate repo and set it to reference local.

I have some stored in my cache, so I haven’t verified it yet.