ubuntu14.04安裝caffe+python介面
1 ubuntu系統安裝
2 caffe環境安裝
General dependencies
sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler sudo apt-get install --no-install-recommends libboost-all-dev
BLAS: sudo apt-get install libatlas-base-dev
Python: sudo apt-get install python-dev python-pip
ubuntu 14.04 :sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev
CUDA: 使用的cuda7.5版本
sudo dpkg -i cuda-repo-ubuntu1404-7-5-local_7.5-18_amd64.deb sudo apt-get update sudo apt-get install -y cuda設定cuda環境:
在/etc/profile檔案最後新增CUDA環境變數:
sudo gedit /etc/profile
PATH=/usr/local/cuda/bin:$PATH
export PATH
儲存後, 執行下列命令, 使環境變數立即生效
source /etc/profile
同時需要新增lib庫路徑: 在 /etc/ld.so.conf.d/加入檔案 cuda.conf
sudo gedit /etc/ld.so.conf.d/cuda.conf
內容如下:
/usr/local/cuda/lib64
儲存後,執行下列命令使之立刻生效
sudo ldconfig
CUDNN:使用4.0版本
tar -zxvf cudnn-7.0-linux-x64-v4.0-prod.tgz cd cuda sudo cp lib64/lib* /usr/local/cuda/lib64/ sudo cp include/cudnn.h /usr/local/cuda/include/ cd /usr/local/cuda/lib64/ sudo chmod +r libcudnn.so.4.0.7 sudo ln -sf libcudnn.so.4.0.7 libcudnn.so.4 sudo ln -sf libcudnn.so.4 libcudnn.so sudo ldconfig
CUDNN:使用5.0版本
tar -zxvf cudnn-7.5-linux-x64-v5.0-ga.tgz cd cuda sudo cp lib64/lib* /usr/local/cuda/lib64/ sudo cp include/cudnn.h /usr/local/cuda/include/ cd /usr/local/cuda/lib64/ sudo chmod +r libcudnn.so.5.0.5 sudo ln -sf libcudnn.so.5.0.5 libcudnn.so.5 sudo ln -sf libcudnn.so.5 libcudnn.so sudo ldconfig
3 python 介面編譯
安裝git和下載caffe原始碼:
sudo apt-get install build-essential cmake git pkg-config git clone https://github.com/BVLC/caffe.git
下載python依賴環境
sudo su sudo apt-get install gfortran pip install --upgrade pip for req in $(cat requirements.txt); do pip install $req; done // 推薦使用國內映象安裝,你懂的,速度快 for req in $(cat requirements.txt); do pip install $req -i http://pypi.douban.com/simple; donefor req in $(cat requirements.txt); do pip install $req -i https://pypi.tuna.tsinghua.edu.cn/simple; done
編譯:先拷貝一份Makefile.config檔案,取消一下注釋
cp Makefile.config.example Makefile.config# cuDNN acceleration switch (uncomment to build with cuDNN).
USE_CUDNN := 1# Uncomment to support layers written in Python (will link against Python libs)
WITH_PYTHON_LAYER := 1使用如下命令編譯caffe和python介面
make -j
make py編譯 pycaffe時報錯:fatal error: numpy/arrayobject.h沒有那個檔案或目錄
其實numpy已經是安裝的,anaconda2裡面有,Python中import numpy也沒有問題,但就是在此處報錯,解決方法:
sudo apt-get install python-numpy
4 使用spyder呼叫python介面和kdevelop呼叫c++
安裝
sudo pip install spyder sudo apt-get install kdevelop