BasisNet: Two-stage Model Synthesis for Efficient Inference

05/07/2021
by   Mingda Zhang, et al.
14

In this work, we present BasisNet which combines recent advancements in efficient neural network architectures, conditional computation, and early termination in a simple new form. Our approach incorporates a lightweight model to preview the input and generate input-dependent combination coefficients, which later controls the synthesis of a more accurate specialist model to make final prediction. The two-stage model synthesis strategy can be applied to any network architectures and both stages are jointly trained. We also show that proper training recipes are critical for increasing generalizability for such high capacity neural networks. On ImageNet classification benchmark, our BasisNet with MobileNets as backbone demonstrated clear advantage on accuracy-efficiency trade-off over several strong baselines. Specifically, BasisNet-MobileNetV3 obtained 80.3 operations, halving the computational cost of previous state-of-the-art without sacrificing accuracy. With early termination, the average cost can be further reduced to 198M MAdds while maintaining accuracy of 80.0

READ FULL TEXT

page 2

page 6

page 8

page 14

research
09/26/2019

Exascale Deep Learning to Accelerate Cancer Research

Deep learning, through the use of neural networks, has demonstrated rema...
research
09/19/2020

ENAS4D: Efficient Multi-stage CNN Architecture Search for Dynamic Inference

Dynamic inference is a feasible way to reduce the computational cost of ...
research
05/28/2018

Sacrificing Accuracy for Reduced Computation: Cascaded Inference Based on Softmax Confidence

We study the tradeoff between computational effort and accuracy in a cas...
research
06/10/2019

FASTER Recurrent Networks for Video Classification

Video classification methods often divide the video into short clips, do...
research
10/31/2022

Tech Report: One-stage Lightweight Object Detectors

This work is for designing one-stage lightweight detectors which perform...
research
12/03/2020

Multiple Networks are More Efficient than One: Fast and Accurate Models via Ensembles and Cascades

Recent work on efficient neural network architectures focuses on discove...
research
04/17/2023

ATHEENA: A Toolflow for Hardware Early-Exit Network Automation

The continued need for improvements in accuracy, throughput, and efficie...

Please sign up or login with your details

Forgot password? Click here to reset