English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Training Convolutional Neural Networks (CNNs) usually requires a large numberof computational resources. In this paper, \textit{SparseTrain} is proposed toaccelerate CNN training by fully exploiting the sparsity. It mainly involvesthree levels of innovations: activation gradients pruning algorithm, sparsetraining dataflow, and accelerator architecture. By applying a stochasticpruning algorithm on each layer, the sparsity of back-propagation gradients canbe increased dramatically without degrading training accuracy and convergencerate. Moreover, to utilize both \textit{natural sparsity} (resulted from ReLUor Pooling layers) and \textit{artificial sparsity} (brought by pruningalgorithm), a sparse-aware architecture is proposed for training acceleration.This architecture supports forward and back-propagation of CNN by adopting1-Dimensional convolution dataflow. We have built %a simple compiler to mapCNNs topology onto \textit{SparseTrain}, and a cycle-accurate architecturesimulator to evaluate the performance and efficiency based on the synthesizeddesign with $14nm$ FinFET technologies. Evaluation results on AlexNet/ResNetshow that \textit{SparseTrain} could achieve about $2.7 \times$ speedup and$2.2 \times$ energy efficiency improvement on average compared with theoriginal training process.

SparseTrain: Exploiting Dataflow Sparsity for Efficient Convolutional Neural Networks Training