SquishedNets: Squishing SqueezeNet further for edge device scenarios via deep evolutionary synthesis
Mohammad Javad Shafiee, Francis Li, Brendan Chwyl, and Alexander Wong

TL;DR
This paper introduces SquishedNets, highly compact deep neural networks derived from SqueezeNet v1.1 through architectural modifications and evolutionary synthesis, optimized for edge devices with limited computational resources.
Contribution
It presents a novel combination of architectural modifications and evolutionary synthesis to create extremely small neural networks tailored for low-resource edge device applications.
Findings
SquishedNets achieve model sizes from 0.95MB to 2.4MB.
They maintain accuracy between 77% and 81.2%.
Processing speeds reach up to 256 images/sec on Nvidia Jetson TX1.
Abstract
While deep neural networks have been shown in recent years to outperform other machine learning methods in a wide range of applications, one of the biggest challenges with enabling deep neural networks for widespread deployment on edge devices such as mobile and other consumer devices is high computational and memory requirements. Recently, there has been greater exploration into small deep neural network architectures that are more suitable for edge devices, with one of the most popular architectures being SqueezeNet, with an incredibly small model size of 4.8MB. Taking further advantage of the notion that many applications of machine learning on edge devices are often characterized by a low number of target classes, this study explores the utility of combining architectural modifications and an evolutionary synthesis strategy for synthesizing even smaller deep neural architectures…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Neural Network Applications · Machine Learning and ELM · Advanced Memory and Neural Computing
Methods*Communicated@Fast*How Do I Communicate to Expedia? · Residual Connection · Convolution · Average Pooling · Fire Module · Global Average Pooling · 1x1 Convolution · Dropout · Xavier Initialization · Max Pooling
