# DeepCABAC: A Universal Compression Algorithm for Deep Neural Networks

## TL;DR

DeepCABAC introduces a novel neural network compression method leveraging video coding techniques, achieving higher compression rates with minimal accuracy loss, exemplified by compressing VGG16 by over 60 times.

## Contribution

It adapts the CABAC video coding technique for neural network compression and proposes a new quantization scheme that balances rate and accuracy.

## Key findings

- Achieves 63.6x compression of VGG16 with no accuracy loss.
- Outperforms previous neural network compression methods.
- Provides open-source implementation for encoding and decoding.

## Abstract

The field of video compression has developed some of the most sophisticated and efficient compression algorithms known in the literature, enabling very high compressibility for little loss of information. Whilst some of these techniques are domain specific, many of their underlying principles are universal in that they can be adapted and applied for compressing different types of data. In this work we present DeepCABAC, a compression algorithm for deep neural networks that is based on one of the state-of-the-art video coding techniques. Concretely, it applies a Context-based Adaptive Binary Arithmetic Coder (CABAC) to the network's parameters, which was originally designed for the H.264/AVC video coding standard and became the state-of-the-art for lossless compression. Moreover, DeepCABAC employs a novel quantization scheme that minimizes the rate-distortion function while simultaneously taking the impact of quantization onto the accuracy of the network into account. Experimental results show that DeepCABAC consistently attains higher compression rates than previously proposed coding techniques for neural network compression. For instance, it is able to compress the VGG16 ImageNet model by x63.6 with no loss of accuracy, thus being able to represent the entire network with merely 8.7MB. The source code for encoding and decoding can be found at https://github.com/fraunhoferhhi/DeepCABAC.