A survey on GAN acceleration using memory compression techniques
Authors:
Open Access
- 19 December 2021
- journal article
- review article
- Published by Springer Nature in Journal of Engineering and Applied Science
- Vol. 68 (1) , 1-23
Abstract
Since its invention, generative adversarial networks (GANs) have shown outstanding results in many applications. GANs are powerful, yet resource-hungry deep learning models. The main difference between GANs and ordinary deep learning models is the nature of their output and training instability. For example, GANs output can be a whole image versus other models detecting objects or classifying images. Thus, the architecture and numeric precision of the network affect the quality and speed of the solution. Hence, accelerating GANs is pivotal. Data transfer is considered the main source of energy consumption, that is why memory compression is a very efficient technique to accelerate and optimize GANs. Two main types of memory compression exist: lossless and lossy ones. Lossless compression techniques are general among all models; thus, we will focus in this paper on lossy techniques. Lossy compression techniques are further classified into (a) pruning, (b) knowledge distillation, (c) low-rank factorization, (d) lowering numeric precision, and (e) encoding. In this paper, we survey lossy compression techniques for CNN-based GANs. Our findings showed the superiority of knowledge distillation over pruning alone and the gaps in the research field that needs to be explored like encoding and different combination of compression techniques.Keywords
All Related Versions
This publication has 31 references indexed in Scilit:
- StackGAN: Text to Photo-Realistic Image Synthesis with Stacked Generative Adversarial Networks2017
- Least Squares Generative Adversarial Networks2017
- Learning Efficient Convolutional Networks through Network Slimming2017
- Unpaired Image-to-Image Translation Using Cycle-Consistent Adversarial Networks2017
- Image-to-Image Translation with Conditional Adversarial Networks2017
- BHNN: A memory-efficient accelerator for compressing deep neural networks with blocked hashing techniques2017
- Speeding up Convolutional Neural Networks with Low Rank Expansions2014
- Sparse Matrices in MATLAB: Design and ImplementationSIAM Journal on Matrix Analysis and Applications, 1992
- Dynamic huffman codingJournal of Algorithms, 1985
- Optimizing a scheme for run length encodingProceedings of the IEEE, 1969