Automated neural architecture search (NAS) methods have been demonstrated as a powerful tool to facilitate neuralarchitecture design. However, the broad applicability of NAS has been restrained due to the difficulty of designingtask-specific search spaces and the necessity and verbosity to implement every NAS component from scratchwhen switching to another search space. In this work, we propose ModularNAS, a framework that implementsessential components of NAS in a modularized and unified manne…