FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable Neural Architecture Search

Abstract

Designing accurate and efficient ConvNets for mobile devices is challenging because the design space is combinatorially large. Due to this, previous neural architecture search (NAS) methods are computationally expensive. ConvNet architecture optimality depends on factors such as input resolution and target devices. However, existing approaches are too expensive for case-by-case redesigns. Also, previous work focuses primarily on reducing FLOPs, but FLOP count does not always reflect actual latency. To address these, we propose a differentiable neural architecture search (DNAS) framework that uses gradient-based methods to optimize ConvNet architectures, avoiding enumerating and training individual architectures separately as in previous methods. FBNets, a family of models discovered by DNAS surpass state-of-the-art models both designed manually and generated automatically. FBNet-B achieves 74.1% top-1 accuracy on ImageNet with 295M FLOPs and 23.1 ms latency on a Samsung S8 phone, 2.4x smaller and 1.5x faster than MobileNetV2-1.3 with similar accuracy. Despite higher accuracy and lower latency than MnasNet, we estimate FBNet-B's search cost is 420x smaller than MnasNet's, at only 216 GPU-hours. Searched for different resolutions and channel sizes, FBNets achieve 1.5% to 6.4% higher accuracy than MobileNetV2. The smallest FBNet achieves 50.2% accuracy and 2.9 ms latency (345 frames per second) on a Samsung S8. Over a Samsung-optimized FBNet, the iPhone-X-optimized model achieves a 1.4x speedup on an iPhone X.

References in corpus (15)

• Adam: A Method for Stochastic Optimization
• Very Deep Convolutional Networks for Large-Scale Image Recognition
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
• Neural Architecture Search with Reinforcement Learning
• Categorical Reparameterization with Gumbel-Softmax
• AMC: AutoML for Model Compression and Acceleration on Mobile Devices
• The Concrete Distribution: A Continuous Relaxation of Discrete Random Variables
• Efficient Neural Architecture Search via Parameter Sharing
• SNAS: Stochastic Neural Architecture Search
• ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
• Mixed Precision Quantization of ConvNets via Differentiable Neural Architecture Search
• Progressive Neural Architecture Search
• Shift: A Zero FLOP, Zero Parameter Alternative to Spatial Convolutions
• SqueezeSegV2: Improved Model Structure and Unsupervised Domain Adaptation for Road-Object Segmentation from a LiDAR Point Cloud

Cited by in corpus (27)

• Single Path One-Shot Neural Architecture Search with Uniform Sampling
• Evaluating the Search Phase of Neural Architecture Search
• Weighted Random Search for CNN Hyperparameter Optimization
• Discovering Neural Wirings
• Sample-Efficient Neural Architecture Search by Learning Action Space
• Memory-Driven Mixed Low Precision Quantization For Enabling Deep Network Inference On Microcontrollers
• How to Train Your Super-Net: An Analysis of Training Heuristics in Weight-Sharing NAS
• HALOC: Hardware-Aware Automatic Low-Rank Compression for Compact Neural Networks
• Hardware/Software Co-Exploration of Neural Architectures
• Anytime Inference with Distilled Hierarchical Neural Ensembles
• Densely Connected Search Space for More Flexible Neural Architecture Search
• BETANAS: BalancEd TrAining and selective drop for Neural Architecture Search
• Single-Path NAS: Device-Aware Efficient ConvNet Design
• Real-Time Execution of Large-scale Language Models on Mobile
• MSG-Transformer: Exchanging Local Spatial Information by Manipulating Messenger Tokens
• RPR: Random Partition Relaxation for Training; Binary and Ternary Weight Neural Networks
• Improving One-shot NAS by Suppressing the Posterior Fading
• Evolving Search Space for Neural Architecture Search
• Neural Architecture Refinement: A Practical Way for Avoiding Overfitting in NAS
• Intra-Ensemble in Neural Networks
• FNA++: Fast Network Adaptation via Parameter Remapping and Architecture Search
• Device-Circuit-Architecture Co-Exploration for Computing-in-Memory Neural Accelerators
• Efficient Novelty-Driven Neural Architecture Search
• DeepSquare: Boosting the Learning Power of Deep Convolutional Neural Networks with Elementwise Square Operators
• GPNAS: A Neural Network Architecture Search Framework Based on Graphical Predictor
• AutoFL: Enabling Heterogeneity-Aware Energy Efficient Federated Learning
• Joint Learning of Neural Transfer and Architecture Adaptation for Image Recognition