深度学习框架 Digits 3.0 安装运行

来源:互联网 发布:淘宝美工教程视频 编辑:程序博客网 时间:2024/06/05 05:06

NVIDIA 不愧是推动Deep learning 的中坚力量,之前运行2.0版本正得心应手时,就推出了3.0 版本。Github地址:https://github.com/NVIDIA/DIGITS,3.0版本的安装使用更为简洁,极易上手。

更新内容:

The new DIGITS 3 release improves training productivity with enhanced workflows.

  • Train neural network models with Torch support (preview), including pre-built AlexNet and GoogleNet models that make it easy to get started.
  • Quickly identify the best model through focused design iteration using the new results browser.
  • Easily manage multiple training jobs to optimize use of system resources.
  • Explore image datasets using the new dataset browser contributed to the DIGITS open source project by Deepomatic.
操作系统: Ubuntu 14.04 64bit.
首先需要配置Caffe 没有配置的同学自行官网。

命令行:
CUDA_REPO_PKG=cuda-repo-ubuntu1404_7.5-18_amd64.deb &&
    wget http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1404/x86_64/$CUDA_REPO_PKG &&
    sudo dpkg -i $CUDA_REPO_PKG

ML_REPO_PKG=nvidia-machine-learning-repo_4.0-2_amd64.deb &&
    wget http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1404/x86_64/$ML_REPO_PKG &&
    sudo dpkg -i $ML_REPO_PKG
待添加完后同样命令行:
apt-get update
apt-get install digits
泡杯Caffe,然后等待安装完成。

安装完成后,在浏览器中输入:
http://localhost/
就可以见到亲切的DIGITS界面了,然后进行train吧。

PS:DIGITS开机自启动,同在一个局域网的机器也可以远程操作。同样在浏览器中输入 运行DIGITS的主机IP地址,同样进入。Amazing~



0 0
原创粉丝点击