Login | Register

StarGAN-v2 compression using knowledge distillation

Title:

StarGAN-v2 compression using knowledge distillation

Kapoor, Paras (2021) StarGAN-v2 compression using knowledge distillation. Masters thesis, Concordia University.

[thumbnail of Kapoor_MCompSc_S2021.pdf]
Preview
Text (application/pdf)
Kapoor_MCompSc_S2021.pdf - Accepted Version
Available under License Spectrum Terms of Access.
10MB

Abstract

Image-to-image translation is used in a broad variety of machine vision and computer graphics applications. These involve mapping grey-scale images to RGB images, deblurring of images, style transfer, transfiguring objects, to name a couple. In addressing complex image-to-image translation issues, Generative Adversarial Networks (GANs) are at the forefront. StarGAN-v2 is a state of the art method for multi-modal multi-domain image-to-image translation that produces different images from a single input image over multiple domains. However, at a parameter count of more than 50M, StarGAN-v2 has a computation bottleneck and consumes more than 60G MACs (Multiply-Accumulate Operations to calculatecomputation expense, 1 MAC = 2 FLOPs) to create one 256×256 image, preventing its widespread adoption. This thesis focuses on the task of compressing StarGAN-v2 using knowledge distillation. Using depthwise separable convolutional layers and reduced channels for intermediate layers, we develop efficient architectures for different StarGAN-v2 modules. In a GAN mini-max optimization environment, the efficient networks are trained with a combination of different distillation losses along with the original objective of StarGAN-v2. Without losing image quality, we reduce the size of the original framework by more than 20× and the computation requirement by more than 5×. The feasibility of the proposed approach is demonstrated by experiments on CelebA-HQ and AFHQ datasets.

Divisions:Concordia University > Gina Cody School of Engineering and Computer Science > Computer Science and Software Engineering
Item Type:Thesis (Masters)
Authors:Kapoor, Paras
Institution:Concordia University
Degree Name:M. Comp. Sc.
Program:Computer Science
Date:29 March 2021
Thesis Supervisor(s):Bui, Tien D.
ID Code:988175
Deposited By: PARAS KAPOOR
Deposited On:29 Jun 2021 22:32
Last Modified:29 Jun 2021 22:32
All items in Spectrum are protected by copyright, with all rights reserved. The use of items is governed by Spectrum's terms of access.

Repository Staff Only: item control page

Downloads per month over past year

Research related to the current document (at the CORE website)
- Research related to the current document (at the CORE website)
Back to top Back to top