给树莓派3安装深度学习Caffe并进行梦境模拟

本文只针对Pi 3,在Pi 2和Pi B+下都不能用,因为涉及很多硬件特性.
系统是2016-09-23-raspbian-jessie.
一开始先sudo apt-get update,然后sudo apt-get upgrade,保证一下系统最新.但是不需要rpi-update.

然后依次安装这些软件:

sudo apt-get install -y gfortran cython
sudo apt-get install -y libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler git
sudo apt-get install --no-install-recommends libboost-all-dev
sudo apt-get install -y python-dev libgflags-dev libgoogle-glog-dev liblmdb-dev libatlas-base-dev python-skimage
sudo pip install pyzmq jsonschema pillow numpy scipy ipython jupyter pyyaml

特别是pip安装时候,因为source.list没在国内有节点,所以慢啊慢.

如果一次不成功多试几次哦.然后再用户目录下创建个目录来装东西.

cd ~
mkdir deepdream
cd deepdream
git clone https://github.com/BVLC/caffe
cd caffe
cp Makefile.config.example Makefile.config
sudo vim<br> Makefile.config

这个git虽然不大,但是在国内嘛,下载还是挺费力.编辑文件打开这个开关.

这个改成/usr/local.

还有这两处.

具体来说,原始:

# CPU_ONLY := 1
/usr/lib/python2.7/dist-packages/numpy/core/include
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib

改成:

CPU_ONLY := 1
/usr/local/lib/python2.7/dist-packages/numpy/core/include
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial/
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/arm-linux-gnueabihf/hdf5/serial/

然后开始编译.

make all
make test
make runtest
make pycaffe
./scripts/download_model_binary.py models/bvlc_googlenet
sudo vim ~/.bashrc

在bashrc里面加入这个.

export PYTHONPATH=/home/pi/deepdream/caffe/python:$PYTHONPATH

还要安装protobuf.

cd ~/deepdream
wget https://github.com/google/protobuf/releases/download/v2.6.1/protobuf-2.6.1.tar.gz
tar xf protobuf-2.6.1.tar.gz
cd protobuf-2.6.1
sudo ./configure
sudo make
sudo make install
sudo nano ~/.bashrc

然后在bashrc里面写这一句.

export LD_LIBRARY_PATH=/usr/local/lib

执行安装:

. ~/.bashrc
cd python
python setup.py build
python setup.py google_test
sudo python setup.py install

可以安装这个来做梦境生成测试.

cd ~/deepdream
git clone https://github.com/JoBergs/deepdream
cd deepdream

因为Pi内存小,比如这个.

python deepdreaming.py -s sky_small.jpg -r

他的梦境是这样的.

自己上传图片做梦也可以,要时刻注意小心不够内存崩溃.比如我拿自己喝水的一个杯子来看.

开始冥想,然后报告了一堆错误.不过还会继续.

看到内存都SWAP都用上了.马上要没资源了.

然后果然报错了.于是修改/etc/dphys-swapfile增加SWAP内存.开始上岸.

友情提示,我这张图片要1.5G的SWAP.刚开始给1G还不行.而且冥想很费时间.程序一开,RAM用光.

随着时间慢慢推移,RAM会越来越用的多.

然后发现怎么觉得多了个人脸.

如果增加到足够多得SWAP,就可以去做更深入得梦境.