# ViP: Virtual Pooling for Accelerating CNN-based Image Classification and Object Detection

## TL;DR

ViP introduces a model-level technique to accelerate CNN-based image classification and object detection, significantly reducing computation time and energy use while maintaining high accuracy, applicable across various models and platforms.

## Contribution

The paper proposes Virtual Pooling (ViP), a novel method that improves CNN speed and energy efficiency with a provable error bound, and demonstrates its effectiveness across multiple models and datasets.

## Key findings

- 2.1x speedup in ImageNet classification with less than 1.5% accuracy loss
- 1.8x speedup in object detection with minimal mAP degradation
- up to 70% reduction in mobile energy consumption

## Abstract

In recent years, Convolutional Neural Networks (CNNs) have shown superior capability in visual learning tasks. While accuracy-wise CNNs provide unprecedented performance, they are also known to be computationally intensive and energy demanding for modern computer systems. In this paper, we propose Virtual Pooling (ViP), a model-level approach to improve speed and energy consumption of CNN-based image classification and object detection tasks, with a provable error bound. We show the efficacy of ViP through experiments on four CNN models, three representative datasets, both desktop and mobile platforms, and two visual learning tasks, i.e., image classification and object detection. For example, ViP delivers 2.1x speedup with less than 1.5% accuracy degradation in ImageNet classification on VGG-16, and 1.8x speedup with 0.025 mAP degradation in PASCAL VOC object detection with Faster-RCNN. ViP also reduces mobile GPU and CPU energy consumption by up to 55% and 70%, respectively. As a complementary method to existing acceleration approaches, ViP achieves 1.9x speedup on ThiNet leading to a combined speedup of 5.23x on VGG-16. Furthermore, ViP provides a knob for machine learning practitioners to generate a set of CNN models with varying trade-offs between system speed/energy consumption and accuracy to better accommodate the requirements of their tasks. Code is available at https://github.com/cmu-enyac/VirtualPooling.