InvSR
InvSR enhances images using a deep noise predictor and pre-trained diffusion models, offering superior results in just one to five steps, ideal for researchers, professionals, and anyone needing high-quality image upsampling.
What is InvSR?
InvSR is an advanced image super-resolution technique that leverages large pre-trained diffusion models to enhance image quality. It uses a partial noise prediction strategy and a deep noise predictor to generate high-resolution images with flexible sampling steps, offering superior performance even in single-step sampling.