Description
ProxylessNAS directly learns neural network architectures on the target task and target hardware without any proxy task. Additional contributions include:
- Using a new path-level pruning perspective for neural architecture search, showing a close connection between NAS and model compression. Memory consumption is saved by one order of magnitude by using path-level binarization.
- Using a novel gradient-based approach (latency regularization loss) for handling hardware objectives (e.g. latency). Given different hardware platforms: CPU/GPU/Mobile, ProxylessNAS enables hardware-aware neural network specialization that’s exactly optimized for the target hardware.
Papers Using This Method
Lightweight Neural Architecture Search for Temporal Convolutional Networks at the Edge2023-01-24NAS-LID: Efficient Neural Architecture Search with Local Intrinsic Dimension2022-11-23Search Space Adaptation for Differentiable Neural Architecture Search in Image Classification2022-06-05AGNAS: Attention-Guided Micro- and Macro-Architecture Search2022-06-05Generalizing Few-Shot NAS with Gradient Matching2022-03-29Towards Tailored Models on Private AIoT Devices: Federated Direct Neural Architecture Search2022-02-23Data-Free Neural Architecture Search via Recursive Label Calibration2021-12-03Profiling Neural Blocks and Design Spaces for Mobile Neural Architecture Search2021-09-25Federated Model Search via Reinforcement Learning2021-07-07Generative Adversarial Neural Architecture Search2021-05-19Generalized Latency Performance Estimation for Once-For-All Neural Architecture Search2021-01-04Efficient Neural Architecture Search for End-to-end Speech Recognition via Straight-Through Gradients2020-11-11How Does Supernet Help in Neural Architecture Search?2020-10-16Angle-based Search Space Shrinking for Neural Architecture Search2020-04-28UNAS: Differentiable Architecture Search Meets Reinforcement Learning2019-12-16Neural Predictor for Neural Architecture Search2019-12-02ProxylessNAS: Direct Neural Architecture Search on Target Task and Hardware2018-12-02