Spandrel gives your project support for various PyTorch architectures meant for AI Super-Resolution, restoration, and inpainting. Based on the model support implemented in chaiNNer.
WIP: I just added architecture code + detection for now. There is no call API, so the models can't really be used rn.
The architectures I added so far are VRT, CAIN, and FLAVR. I also wanted to add RIFE, but haven't gotten to that yet, because RIFE is still being actively improved, leading to 10\~15 subtly different variants of RIFE that are all incompatible with each other. I might be able to reduce the number of variants by introducing new hyperparameters, but I still expect to end up with 5\~10 variants.
WIP: I just added architecture code + detection for now. There is no call API, so the models can't really be used rn.
The architectures I added so far are VRT, CAIN, and FLAVR. I also wanted to add RIFE, but haven't gotten to that yet, because RIFE is still being actively improved, leading to 10\~15 subtly different variants of RIFE that are all incompatible with each other. I might be able to reduce the number of variants by introducing new hyperparameters, but I still expect to end up with 5\~10 variants.