LGMLJul 29, 2020

Compressing Deep Neural Networks via Layer Fusion

arXiv:2007.14917v17 citations
Originality Incremental advance
AI Analysis

This addresses the problem of reducing model size for deployment in resource-constrained environments, but it is incremental as it builds on existing compression methods.

The paper tackles model compression by proposing layer fusion, a technique that combines weights of similar layers to reduce network size with minimal computational overhead, achieving up to 3.33x compression on CIFAR-10 with less than 2% accuracy drop and compressing transformers to 20% of original size on WikiText-2 within 5 perplexity points.

This paper proposes \textit{layer fusion} - a model compression technique that discovers which weights to combine and then fuses weights of similar fully-connected, convolutional and attention layers. Layer fusion can significantly reduce the number of layers of the original network with little additional computation overhead, while maintaining competitive performance. From experiments on CIFAR-10, we find that various deep convolution neural networks can remain within 2\% accuracy points of the original networks up to a compression ratio of 3.33 when iteratively retrained with layer fusion. For experiments on the WikiText-2 language modelling dataset where pretrained transformer models are used, we achieve compression that leads to a network that is 20\% of its original size while being within 5 perplexity points of the original network. We also find that other well-established compression techniques can achieve competitive performance when compared to their original networks given a sufficient number of retraining steps. Generally, we observe a clear inflection point in performance as the amount of compression increases, suggesting a bound on the amount of compression that can be achieved before an exponential degradation in performance.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes