StableSR v.s. Stable Diffusion x4 Upscaler

[XuejiFang]

Thanks for your excellent work, which has inprised me a lot.

However, I wonder if you had compared the StableSR proposed in your paper with Stable Diffusion x4 Upscaler proposed by StabilityAI. While the SD Upscaler requires training from scratch, once it has been trained, how does its generalization performance in other domains compare to your StableSR?

[IceClear]

Please refer to the paper.