NN-Stretch: Automatic Neural Network Branching for Parallel Inference on Heterogeneous Multi-Processors
Jianyu Wei, Ting Cao, Shijie Cao, Shiqi Jiang, Shaowei Fu, Mao Yang, Yanyong Zhang, Yunxin Liu
Abstract
Mobile devices are increasingly equipped with heterogeneous multiprocessors, e.g., CPU + GPU + DSP. Yet existing Neural Network (NN) inference fails to fully utilize the computing power of the heterogeneous multi-processors due to the sequential structures of NN models. Towards this end, this paper proposes NN-Stretch, a new model adaption strategy, as well as the supporting system. It automatically branches a given model according to the processor architecture characteristics. Compared to other popular model adaption techniques such as model pruning that often sacrifices accuracy, NN-Stretch accelerates inference while preserving accuracy.
Topics & Concepts
Computer sciencePruningInferenceBranching (polymer chemistry)Artificial neural networkDigital signal processingDeep neural networksParallel computingComputer architectureArtificial intelligenceComputer hardwareAgronomyMaterials scienceBiologyComposite materialAdvanced Neural Network ApplicationsNeural Networks and ApplicationsParallel Computing and Optimization Techniques