TypeError: RegionalDiffusionXLPipeline.__init__() missing 2 required positional arguments: 'text_encoder_2' and 'tokenizer_2'

[hq0709]

TypeError: RegionalDiffusionXLPipeline.init() missing 2 required positional arguments: 'text_encoder_2' and 'tokenizer_2'

[ZhaochenYu0201]

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.

[hq0709]

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!

[ZhaochenYu0201]

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.

[hq0709]

Thanks for the explanation :)