用户可利用该项目实现个性化边缘DNN模型的优化。OCAP通过输入图像的中间激活识别并移除类无关通道,提供CIFAR和ImageNet数据集上的模型剪枝代码,含MobileNetV2、ResNet、Vgg等模型定义及辅助函数。【此简介由AI生成】
| Files | Last commit | Last update |
|---|---|---|
| 10 months ago | ||
| 10 months ago | ||
| 10 months ago | ||
| 10 months ago | ||
| 10 months ago | ||
| 10 months ago | ||
| 10 months ago |
OCAP: On-device Class-Aware Pruning for personalized edge DNN models
This repository contains the code for OCAP.
OCAP proposes a new class-aware pruning method based on the intermediate activation of input images to identify and remove the class-irrelevant channels.
For other details, please refer to the paper.
All the model definitions are available under ./models,
and the auxiliary functions are available under ./utils.
OCAP on CIFAR
MobileNetV2_Cifar_Pruning.py
ResNet_Cifar_Pruning.py
Vgg_Cifar_Pruning.py
OCAP on ImageNet
Vgg_ImageNet_Pruning.py
Introduction
用户可利用该项目实现个性化边缘DNN模型的优化。OCAP通过输入图像的中间激活识别并移除类无关通道,提供CIFAR和ImageNet数据集上的模型剪枝代码,含MobileNetV2、ResNet、Vgg等模型定义及辅助函数。【此简介由AI生成】
Customize your domain