Abstract: Transformer-like network has shown remarkable high performance in both natural language processing and computer vision. However, the huge computational demands in non-linear floating-point ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results