Skip to content

Alii-Ganjj/InterpretationsSteeredPruning

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 

Repository files navigation

Interpretations Steered Network Pruning via Amortized Inferred Saliency Maps - ECCV 2022

This repository contains the official PyTorch implementation for the paper:

Interpretations Steered Network Pruning via Amortized Inferred Saliency Maps
Alireza Ganjdanesh*, Shangqian Gao*, Heng Huang
University of Pittsburgh
ECCV 2022

ISP Framework

Abstract

Convolutional Neural Networks (CNNs) compression is crucial to deploying these models in edge devices with limited resources. Existing channel pruning algorithms for CNNs have achieved plenty of success on complex models. They approach the pruning problem from various perspectives and use different metrics to guide the pruning process. However, these metrics mainly focus on the model’s ‘outputs’ or ‘weights’ and neglect its ‘interpretations’ information. To fill in this gap, we propose to address the channel pruning problem from a novel perspective by leveraging the interpretations of a model to steer the pruning process, thereby utilizing information from both inputs and outputs of the model. However, existing interpretation methods cannot get deployed to achieve our goal as either they are inefficient for pruning or may predict non-coherent explanations. We tackle this challenge by introducing a selector model that predicts real-time smooth saliency masks for pruned models. We parameterize the distribution of explanatory masks by Radial Basis Function (RBF)-like functions to incorporate geometric prior of natural images in our selector model’s inductive bias. Thus, we can obtain compact representations of explanations to reduce the computational costs of our pruning method. We leverage our selector model to steer the network pruning by maximizing the similarity of explanatory representations for the pruned and original models. Extensive experiments on CIFAR-10 and ImageNet benchmark datasets demonstrate the efficacy of our proposed method.

Repository Structure

As our architectures and implementations for CIFAR-10 and ImageNet are different, we provide our code for CIFAR-10 and ImageNet in two separate branches. You can access them after cloning the project with the following commands:

git checkout CIFAR10
git checkout ImageNet

Citation

If you find this project helpful, please consider citing our paper:

@inproceedings{ganjdanesh2022isp,  
  title={Interpretations Steered Network Pruning via Amortized Inferred Saliency Maps},
  author={Ganjdanesh, Alireza and Gao, Shangqian and Huang, Heng},
  booktitle={Proceedings of the European Conference on Computer Vision (ECCV)},
  year={2022}
  }

Contact

If you have any questions or suggestions, feel free to contact us. (alireza.ganjdanesh@pitt.edu)

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published