Installing cuda 7.0 on Ubuntu 14.04

Categories: cuda

Using python 2.7/ opencv3

Was getting error ‘undefined cv::imread’1 so, I fixed it by editing the Makefile of caffe to add opencv_imgcodecs

LIBRARIES += glog gflags protobuf leveldb snappy \
        lmdb \
        boost_system \
        hdf5_hl hdf5 \
        opencv_imgcodecs opencv_highgui opencv_imgproc opencv_core pthread

https://github.com/BVLC/caffe/issues/2288

2

added line in Makefile.config and commented the previous line INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial/ LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu/hdf5/serial/

https://github.com/BVLC/caffe/issues/2690 Followed the guide from here

https://github.com/BVLC/caffe/issues/1276

export PYTHONPATH=${CAFFE_HOME}/python:$PYTHONPATH

import caffe

CUDNN_STATUS_ARCH_MISMATCH

I solved it commenting both cuDNN and CPU possibilities at the beginning of the Makefile.config file in order to work only with CUDA and then rebuilding the library with the “make clean/all/test/runtest” commands. It shows a message skipping two test samples when running “make test”, since they are the cuDNN based samples that I am not using anymore. Then, I was able run the mnist/train_lenet.sh example.

#tell to use theano echo -e “\n[global]\nfloatX=float32\ndevice=gpu\n[mode]=FAST_RUN\n\n[nvcc]\nfastmath=True\n\n[cuda]\nroot=/usr/local/cuda” » ~/.theanorc

pip install theano

http://markus.com/install-theano-on-aws/

python python -c "import os, theano; print os.path.dirname(theano.__file__)"/misc/check_blas.py