NVIDIA DIGITS Natively 安装

按照这个网页做https://github.com/dusty-nv/jetson-inference/blob/master/docs/digits-native.md

主要是自己记录一下过程

1. 先安装驱动
2. 安装cudnn
3. 安装nvcaffe
4. 安装Digits

安装驱动

$ sudo apt-get install nvidia-384	# use nvidia-375 for alternate version
$ sudo reboot

查看驱动安装情况：lsmod

$ lsmod | grep nvidia
nvidia_uvm            647168  0
nvidia_drm             49152  1
nvidia_modeset        790528  4 nvidia_drm
nvidia              12144640  60 nvidia_modeset,nvidia_uvm
drm_kms_helper        167936  1 nvidia_drm
drm                   368640  4 nvidia_drm,drm_kms_helper

测试cuda

$ cd /usr/local/cuda/samples
$ sudo make
$ cd bin/x86_64/linux/release/
$ ./deviceQuery
$ ./bandwidthTest --memory=pinned

安装cudnn

$ sudo dpkg -i libcudnn<version>_amd64.deb
$ sudo dpkg -i libcudnn-dev_<version>_amd64.deb

安装nvcaffe

$ git clone -b caffe-0.15 https://github.com/NVIDIA/caffe

mkdir build
cd build
cmake ..
make all
make install
make runtest

或者使用make

cp Makefile.config.example Makefile.config
# Adjust Makefile.config (for example, if using Anaconda Python, or if cuDNN is desired)
make all
make test
make runtest

然后把以下环境变量加入到 ~/.bashrc

export CAFFE_ROOT=/home/sl/DIGITS/caffe/build
export PYTHONPAH=/home/sl/DIGITS/caffe/build/python:$PYTHONPATH

安装Digits

git clone https://github.com/nvidia/DIGITS

运行digits

在digits目录下面运行，digits/jobs下面保存数据集和model

➜  digits git:(master) ✗ ./digits-devserver 
  ___ ___ ___ ___ _____ ___
 |   \_ _/ __|_ _|_   _/ __|
 | |) | | (_ || |  | | \__ \
 |___/___\___|___| |_| |___/ 6.1.1