We provide the code for MLoRE, a new multi-task framework for dense prediction. Our code is implemented on PASCAL-Context and NYUD-v2 based on ViT. MLoRE builds a novel decoder module based on ...
Non-autoregressive Transformer (NAT) is an efficient type of image synthesis model. However, their performance is generally inferior to state-of-the-art image generation models (e.g. diffusion models) ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results