You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thank you for your feedback, but the information you have provided is limited, please give more detailed configuration so that we can help you.
For now, considering the error message, i think that is because you use RegionalDiffusionXLPipeline which only supports SDXL models. I suggest you to check if you used sdv1.4/1.5/2.0/2.1 models , if so, you can use RegionalDiffusionPipeline which is compatible with these models for generation.
Thanks for your reply. I have this message when I try to utilize the anything-v3 model, as well as the absolutereality model. Should I change the pipeline? Thanks!
Anything-v3 is SDv1.4/1.5 based architecture, so you can use RegionalDiffusionPipeline to load the model. The reason why you got this error is because SDXL model have an extra textencoder and an extra tokenizer compare to SD v1.4/1.5 models, so if you use RegionalDiffusionXLPipeline it will report the model is missing some components. I just update the REAMDE to clarify the issuse, thank you for your gentle remind.
TypeError: RegionalDiffusionXLPipeline.init() missing 2 required positional arguments: 'text_encoder_2' and 'tokenizer_2'
The text was updated successfully, but these errors were encountered: