从VGG提取功能 [英] Extracting Features from VGG

问题描述

我想使用微调的VGG-19网络从MS COCO数据集的图像中提取特征。

I want to extract features from images in MS COCO dataset using a fine-tuned VGG-19 network.

但是，每个图像需要大约6〜7秒，每1k图像大约2小时。

However, it takes about 6~7 seconds per image, roughly 2 hours per 1k images. (even longer for other fine-tuned models)

MS COCO数据集中有120k张图片，因此至少需要10天。

There are 120k images in MS COCO dataset, so it'll take at least 10 days.

有什么方法可以加快特征提取过程吗？

Is there any way that I can speed up the feature extraction process?

推荐答案

这不只是一个命令。首先，你必须检查你的GPU是否足够强大，以克服深的CNNs。了解你的GPU模型可以回答这个问题。

Well, this is not just a command. First you must check whether your GPU is powerful enough to wrestle with deep CNNs. Knowing your GPU model can answer this question.

其次，你必须编译和构建Caffe框架与CUDA和GPU启用（CPU_Oly禁用）在Makefile.config或CMakeLists.txt）。

Second, you have to compile and build Caffe framework with CUDA and GPU-enabled (CPU_Only disabled) in the Makefile.config (or CMakeLists.txt).

通过所有必要步骤（安装Nvidia驱动程序，安装CUDA等），您可以构建caffe以供GPU使用。

Passing all required steps (installing Nvidia Driver, installing CUDA and etc.) you can build caffe for GPU-use. Then by passing the GPU_Device_ID in your command-line you can benefit from speed provided by them.

按照 this 链接用于使用GPU构建Caffe。

Follow this link for building Caffe using GPU.

希望它有帮助

这篇关于从VGG提取功能的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！