HardCoRe-NAS: Hard Constrained diffeRentiable Neural Architecture Search
arXiv:2102.11646
Abstract
Realistic use of neural networks often requires adhering to multiple constraints on latency, energy and memory among others. A popular approach to find fitting networks is through constrained Neural Architecture Search (NAS), however, previous methods enforce the constraint only softly. Therefore, the resulting networks do not exactly adhere to the resource constraint and their accuracy is harmed. In this work we resolve this by introducing Hard Constrained diffeRentiable NAS (HardCoRe-NAS), that is based on an accurate formulation of the expected resource requirement and a scalable search method that satisfies the hard constraint throughout the search. Our experiments show that HardCoRe-NAS generates state-of-the-art architectures, surpassing other NAS methods, while strictly satisfying the hard resource constraints without any tuning required.
Niv Nayman and Yonathan Aflalo contributed equally. An implementation of HardCoRe-NAS is available at: https://github.com/Alibaba-MIIL/HardCoReNAS
References in corpus (6)
- PyTorch: An Imperative Style, High-Performance Deep Learning Library
- Distilling the Knowledge in a Neural Network
- Neural Architecture Search with Reinforcement Learning
- SMASH: One-Shot Model Architecture Search through HyperNetworks
- TF-NAS: Rethinking Three Search Freedoms of Latency-Constrained Differentiable Neural Architecture Search
- Global Convergence of Frank Wolfe on One Hidden Layer Networks