Will it improve performance if I apply other stable diffusion model which is finetuned on certain content images?

[eeyrw]

As title said.

[0x3f3f3f3fun]

I think Stable Diffusion v2 may not be suitable for some special domains, such as medical image. The reason is that stable diffusion is not able to generate those domain-specific images. So if you use other diffusion models finetuned on certain content images, I guess🤔 the performance will be improved.

[eeyrw]

Yes. The certain content images refer general images with specific content such as human face rather than domain specific images. Is replacing weight of UNet by existing model and not finetuning LAControNet (plug-and-play) OKay?