DeepRebirth: Accelerating Deep Neural Network Execution on Mobile Devices

TL;DR

DeepRebirth introduces a novel framework that accelerates deep neural network execution on mobile devices by slimming non-tensor layers, achieving over 3x speed-up with minimal accuracy loss.

Contribution

The paper proposes a layer slimming method for non-tensor layers in neural networks, significantly improving runtime speed and memory efficiency on mobile devices.

Findings

Over 3x speed-up on GoogLeNet with minimal accuracy drop

Reduces runtime memory by 2.5x

Achieves 65ms inference on Samsung Galaxy S6 CPU

Abstract

Deploying deep neural networks on mobile devices is a challenging task. Current model compression methods such as matrix decomposition effectively reduce the deployed model size, but still cannot satisfy real-time processing requirement. This paper first discovers that the major obstacle is the excessive execution time of non-tensor layers such as pooling and normalization without tensor-like trainable parameters. This motivates us to design a novel acceleration framework: DeepRebirth through "slimming" existing consecutive and parallel non-tensor and tensor layers. The layer slimming is executed at different substructures: (a) streamline slimming by merging the consecutive non-tensor and tensor layer vertically; (b) branch slimming by merging non-tensor and tensor branches horizontally. The proposed optimization operations significantly accelerate the model execution and also greatly reduce the run-time memory cost since the slimmed model architecture contains less hidden layers. To maximally avoid accuracy loss, the parameters in new generated layers are learned with layer-wise fine-tuning based on both theoretical analysis and empirical verification. As observed in the experiment, DeepRebirth achieves more than 3x speed-up and 2.5x run-time memory saving on GoogLeNet with only 0.4% drop of top-5 accuracy on ImageNet. Furthermore, by combining with other model compression techniques, DeepRebirth offers an average of 65ms inference time on the CPU of Samsung Galaxy S6 with 86.5% top-5 accuracy, 14% faster than SqueezeNet which only has a top-5 accuracy of 80.5%.