With the flourish of differentiable neural architecture search (NAS),
automatically searching latency-constrained architectures gives a new
perspective to reduce human labor and expertise. However, the searched
architectures are usually suboptimal in accuracy and may have large jitters
around the target latency. In this paper, we rethink three freedoms of
differentiable NAS, i.e. operation-level, depth-level and width-level, and
propose a novel method, named Three-Freedom NAS (TF-NAS), to achieve both good
classification accuracy and precise latency constraint. For the
operation-level, we present a bi-sampling search algorithm to moderate the
operation collapse. For the depth-level, we introduce a sink-connecting search
space to ensure the mutual exclusion between skip and other candidate
operations, as well as eliminate the architecture redundancy. For the
width-level, we propose an elasticity-scaling strategy that achieves precise
latency constraint in a progressively fine-grained manner. Experiments on
ImageNet demonstrate the effectiveness of TF-NAS. Particularly, our searched
TF-NAS-A obtains 76.9% top-1 accuracy, achieving state-of-the-art results with
less latency. The total search time is only 1.8 days on 1 Titan RTX GPU. Code
is available at this https URL

本文通过重新思考可微分 NAS 的三个自由度：操作级，深度级和宽度级，提出了一种新的方法，名为 Three-Freedom NAS（TF-NAS），以实现良好的分类准确性和精确的延迟约束。该方法还包括双采样搜索算法、汇流连接搜索空间和弹性缩放策略。在 ImageNet 实验中证明了 TF-NAS 的有效性，尤其是 TF-NAS-A 在 1.8 天内只用 1 个 Titan RTX GPU 时间就获得了 76.9％的顶部 1 准确度，实现了最新的成果。