From the Makers
of SqueezeNet

Research

Tips and Tricks for Developing Smaller Neural Nets

Forrest Iandola

Invited Talk at the CVPR Workshop on Efficient Deep Learning for Computer Vision, 2018

DeepScale

Watch

Co-Design of Deep Neural Nets and Neural Net Accelerators for Embedded Vision Applications

Alon Amid, Krste Asanovic, Amir Gholami, Kurt Keutzer, Kiseok Kwon, Bichen Wu.

Invited Talk at the CVPR Workshop on Efficient Deep Learning for Computer Vision, 2018

UC Berkeley, DeepScale, and Samsung Research

Watch

How Autonomous Driving Challenges Computer Vision Research

Kurt Keutzer, Alon Amid, Krste Asanovic, Amir Gholami, Peter Jin, Kiseok Kwon, Alvin Wan, Bichen Wu, Xiangyu Yue, Sichen Zhao, Forrest Iandola, Ben Landen, Paden Tomasello

Invited Talk at the CVPR Workshop on Autonomous Driving, 2018

UC Berkeley, DeepScale & Samsung Research

Watch

How to become a Full-Stack Deep Learning Engineer

Forrest Iandola

Silicon Valley Deep Learning Group, 2018

DeepScale

Watch

Small Deep Neural Networks – Their Advantages, and Their Design

Forrest Iandola and Kurt Keutzer

Invited Talk at the ICML TinyML Workshop, 2017

DeepScale

Watch

A Shallow Dive into Training Deep Neural Networks

Sammy Sidhu

Embedded Vision Summit, 2017

DeepScale

Watch
https://arxiv.org/pdf/1811.07070.pdf

DSCnet: Replicating Lidar Point Clouds with Deep Sensor Cloning

Paden Tomasello, Sammy Sidhu, Anting Shen, Matthew W. Moskewicz, Nobie Redmon, Gayatri Joshi, Romi Phadte, Paras Jain, Forrest Iandola

arXiv technical report, 2018

DeepScale

Read
https://arxiv.org/pdf/1710.02759.pdf

Small Neural Nets Are Beautiful: Enabling Embedded Systems with Small Deep-Neural-Network Architectures

Forrest Iandola, Kurt Keutzer

Keynote at ESWEEK, 2017

DeepScale & UC Berkeley

Read
https://arxiv.org/pdf/1612.01051.pdf

30x speedup for object detection dnns

SqueezeDet: Unified, Small, Low Power Fully Convolutional Neural Networks For Real-time Object Detection For Autonomous Driving

Bichen Wu, Alvin Wan, Forrest Iandola, Peter H. Jin, Kurt Keutzer

CVPR Embedded Vision Workshop, 2017 (Best Paper Award)

DeepScale & UC Berkeley

Read
https://arxiv.org/pdf/1602.07360.pdf

50-500x smaller dnn models for image classification

SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size

Forrest N. Iandola, Song Han, Matthew W. Moskewicz, Khalid Ashraf, William J. Dally, Kurt Keutzer

arXiv technical report, 2016

DeepScale, UC Berkeley & Stanford University

Read

Boda-RTC: Productive Generation of Portable, Efficient Code for Convolutional Neural Networks on Mobile Computing Platforms

Matthew Moskewicz, Forrest Iandola, and Kurt Keutzer

IEEE International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob), 2016

DeepScale and UC Berkeley

Read

Research Grant Awards:

  • NSF Small Business Innovative Research (SBIR) Phase I Grant Award – Dec 2016
  • NSF Small Business Innovative Research (SBIR) Phase II Grant Award – April 2018