bug:convert controlnet model(custom datasets train) to diffusers failure

[chenjingcheng]

I trained a controlnet model based on sd1.5 or sd2.1 using my own data. The trained controlnet model performs inference correctly. However, when I try to convert it using the scripts/convert_original_controlnet_to_diffusers.py script, I encounter the following error:
Traceback (most recent call last):
File "/home/jacky/aiproject/diffusers/scripts/convert_original_controlnet_to_diffusers.py", line 96, in
controlnet = download_controlnet_from_original_ckpt(
File "/home/jacky/anaconda3/envs/diffusers/lib/python3.10/site-packages/diffusers/pipelines/stable_diffusion/convert_from_ckpt.py", line 1557, in download_controlnet_from_original_ckpt
controlnet_model = convert_controlnet_checkpoint(
File "/home/jacky/anaconda3/envs/diffusers/lib/python3.10/site-packages/diffusers/pipelines/stable_diffusion/convert_from_ckpt.py", line 1038, in convert_controlnet_checkpoint
controlnet_model = ControlNetModel(**ctrlnet_config)
File "/home/jacky/anaconda3/envs/diffusers/lib/python3.10/site-packages/diffusers/configuration_utils.py", line 614, in inner_init
init(self, *args, **init_kwargs)
TypeError: ControlNetModel.init() got an unexpected keyword argument 'addition_embed_type'

[patrickvonplaten]

Hey @chenjingcheng,

Can you please provide a reproducible code snippet?

[haoli-zbdbc]

same question!

[patrickvonplaten]

cc @sayakpaul is this related to #4119

Will try to fix later in #4084

[sayakpaul]

Yeah seems to be related.

[patrickvonplaten]

Ok took a look, I think this will be fixed with #4038

[patrickvonplaten]

Should be fixed now - can you check @chenjingcheng ?

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.