HAQ: Hardware-Aware Automated Quantization with Mixed Precision

Abstract

Model quantization is a widely used technique to compress and accelerate deep neural network (DNN) inference. Emergent DNN hardware accelerators begin to support mixed precision (1-8 bits) to further improve the computation efficiency, which raises a great challenge to find the optimal bitwidth for each layer: it requires domain experts to explore the vast design space trading off among accuracy, latency, energy, and model size, which is both time-consuming and sub-optimal. Conventional quantization algorithm ignores the different hardware architectures and quantizes all the layers in a uniform way. In this paper, we introduce the Hardware-Aware Automated Quantization (HAQ) framework which leverages the reinforcement learning to automatically determine the quantization policy, and we take the hardware accelerator's feedback in the design loop. Rather than relying on proxy signals such as FLOPs and model size, we employ a hardware simulator to generate direct feedback signals (latency and energy) to the RL agent. Compared with conventional methods, our framework is fully automated and can specialize the quantization policy for different neural network architectures and hardware architectures. Our framework effectively reduced the latency by 1.4-1.95x and the energy consumption by 1.9x with negligible loss of accuracy compared with the fixed bitwidth (8 bits) quantization. Our framework reveals that the optimal policies on different hardware architectures (i.e., edge and cloud architectures) under different resource constraints (i.e., latency, energy and model size) are drastically different. We interpreted the implication of different quantization policies, which offer insights for both neural network architecture design and hardware architecture design.

CVPR 2019. The first three authors contributed equally to this work. Project page: https://hanlab.mit.edu/projects/haq/

References in corpus (10)

• Continuous control with deep reinforcement learning
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• Neural Architecture Search with Reinforcement Learning
• Binarized Neural Networks: Training Deep Neural Networks with Weights and Activations Constrained to +1 or -1
• DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients
• AMC: AutoML for Model Compression and Acceleration on Mobile Devices
• Quantizing deep convolutional networks for efficient inference: A whitepaper
• Trained Ternary Quantization
• PACT: Parameterized Clipping Activation for Quantized Neural Networks
• Path-Level Network Transformation for Efficient Architecture Search

Cited by in corpus (12)

• TSM: Temporal Shift Module for Efficient Video Understanding
• LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inference
• And the Bit Goes Down: Revisiting the Quantization of Neural Networks
• ReLeQ: A Reinforcement Learning Approach for Deep Quantization of Neural Networks
• Low-Memory Neural Network Training: A Technical Report
• Memory-Driven Mixed Low Precision Quantization For Enabling Deep Network Inference On Microcontrollers
• Towards Unified INT8 Training for Convolutional Neural Network
• GNN is a Counter? Revisiting GNN for Question Answering
• GeneCAI: Genetic Evolution for Acquiring Compact AI
• RPR: Random Partition Relaxation for Training; Binary and Ternary Weight Neural Networks
• MetaMix: Meta-state Precision Searcher for Mixed-precision Activation Quantization
• n-hot: Efficient bit-level sparsity for powers-of-two neural network quantization