Thanks for the inspired work! it shows a very promising way to get perfect recon and accelerated generation training
I'm just wondering is there any plan to release the model weights( including the RAC decoder and the ImageNet DiT train on the RAC feature? Hope to get your reply
Thanks for the inspired work! it shows a very promising way to get perfect recon and accelerated generation training
I'm just wondering is there any plan to release the model weights( including the RAC decoder and the ImageNet DiT train on the RAC feature? Hope to get your reply