Towards Efficient and Data Agnostic Image Classification Training Pipeline for Embedded Systems

TL;DR

This paper proposes a data-agnostic, efficient image classification training pipeline that automatically adjusts hyperparameters and uses lightweight CNN architectures, enabling good performance across diverse datasets without manual tuning.

Contribution

It introduces a training pipeline that automates hyperparameter selection and employs modern lightweight CNNs for robust, data-agnostic image classification on embedded systems.

Findings

Achieves reasonable accuracy across multiple datasets without manual tuning

Uses lightweight CNN architectures suitable for CPU deployment

Provides an open-source implementation within OpenVINO extensions

Abstract

Nowadays deep learning-based methods have achieved a remarkable progress at the image classification task among a wide range of commonly used datasets (ImageNet, CIFAR, SVHN, Caltech 101, SUN397, etc.). SOTA performance on each of the mentioned datasets is obtained by careful tuning of the model architecture and training tricks according to the properties of the target data. Although this approach allows setting academic records, it is unrealistic that an average data scientist would have enough resources to build a sophisticated training pipeline for every image classification task he meets in practice. This work is focusing on reviewing the latest augmentation and regularization methods for the image classification and exploring ways to automatically choose some of the most important hyperparameters: total number of epochs, initial learning rate value and it's schedule. Having a training procedure equipped with a lightweight modern CNN architecture (like bileNetV3 or EfficientNet), sufficient level of regularization and adaptive to data learning rate schedule, we can achieve a reasonable performance on a variety of downstream image classification tasks without manual tuning of parameters to each particular task. Resulting models are computationally efficient and can be deployed to CPU using the OpenVINO toolkit. Source code is available as a part of the OpenVINO Training Extensions (https://github.com/openvinotoolkit/training_extensions).