Abstract: Recently, there have been increasing demands to construct compact deep architectures to remove unnecessary redundancy and to improve the inference speed. While many recent works focus on ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results