Pushing the limits of RNN Compression
This addresses the challenge of efficient RNN deployment for applications on devices with limited resources, representing a strong specific gain rather than a foundational advancement.
The paper tackles the problem of deploying large Recurrent Neural Networks (RNNs) on resource-constrained devices by introducing a compression method using Kronecker product (KP), achieving 16-38x compression with minimal accuracy loss and outperforming state-of-the-art techniques in accuracy and inference speed across multiple benchmarks.
Recurrent Neural Networks (RNN) can be difficult to deploy on resource constrained devices due to their size. As a result, there is a need for compression techniques that can significantly compress RNNs without negatively impacting task accuracy. This paper introduces a method to compress RNNs for resource constrained environments using Kronecker product (KP). KPs can compress RNN layers by 16-38x with minimal accuracy loss. We show that KP can beat the task accuracy achieved by other state-of-the-art compression techniques (pruning and low-rank matrix factorization) across 4 benchmarks spanning 3 different applications, while simultaneously improving inference run-time.