Dockerfile for cudnn based caffe
Installation:
-
Install CUDA-7.5 on the host machine
-
Add cuddn.tgz (https://developer.nvidia.com/cudnn) to project root.
-
Build docker image (
docker run build -t cudnn-caffe
) -
Run docker container with nvidia devices mounted:
docker run -it -v /opt/data:/opt/data --device /dev/nvidiactl --device /dev/nvidia0 cudnn-caffe
4. re-install cuda (following recomedations from http://tleyden.github.io/blog/2015/11/22/running-neural-style-on-an-aws-gpu-instance):
wget http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1404/x86_64/cuda-repo-ubuntu1404_7.5-18_amd64.deb
sudo dpkg -i cuda-repo-ubuntu1404_7.5-18_amd64.deb
sudo apt-get update
sudo apt-get upgrade -y
sudo apt-get install -y opencl-headers build-essential protobuf-compiler \
libprotoc-dev libboost-all-dev libleveldb-dev hdf5-tools libhdf5-serial-dev \
libopencv-core-dev libopencv-highgui-dev libsnappy-dev libsnappy1 \
libatlas-base-dev cmake libstdc++6-4.8-dbg libgoogle-glog0 libgoogle-glog-dev \
libgflags-dev liblmdb-dev git python-pip gfortran
sudo apt-get clean
sudo apt-get install -y linux-image-extra-`uname -r` linux-headers-`uname -r` linux-image-`uname -r`
sudo apt-get install -y cuda
If you encounter the error like this: "error while loading shared libraries: libcudart.so.7.5" run the following commands:
export PATH=$PATH:/usr/local/cuda-7.5/bin
export LD_LIBRARY_PATH=:/usr/local/cuda-7.5/lib64
sudo ldconfig
- Run caffe