Once for All: Train One Network and Specialize it for Efficient Deployment Post date February 12, 2021 ← Aggressive Compression of MobileNets Using Hybrid Ternary Layers → Adaptive Video Sampling for Energy-Efficient Object Detection