EDCompress: Energy-Aware Model Compression for Dataflows

TL;DR

EDCompress is a reinforcement learning-based method that optimizes CNN model compression for various dataflows, significantly reducing energy consumption on edge devices while maintaining accuracy.

Contribution

It introduces a novel energy-aware compression approach tailored for different dataflow types, addressing a gap in existing work.

Findings

Achieves up to 37X energy efficiency improvements.

Effectively finds optimal dataflow types for neural networks.

Maintains negligible accuracy loss during compression.

Abstract

Edge devices demand low energy consumption, cost and small form factor. To efficiently deploy convolutional neural network (CNN) models on edge device, energy-aware model compression becomes extremely important. However, existing work did not study this problem well because the lack of considering the diversity of dataflow types in hardware architectures. In this paper, we propose EDCompress, an Energy-aware model compression method for various Dataflows. It can effectively reduce the energy consumption of various edge devices, with different dataflow types. Considering the very nature of model compression procedures, we recast the optimization process to a multi-step problem, and solve it by reinforcement learning algorithms. Experiments show that EDCompress could improve 20X, 17X, 37X energy efficiency in VGG-16, MobileNet, LeNet-5 networks, respectively, with negligible loss of accuracy. EDCompress could also find the optimal dataflow type for specific neural networks in terms of energy consumption, which can guide the deployment of CNN models on hardware systems.