I realized that compiling MatConvNet in Ubuntu is better to do it from the linux shell as supposed to from matlab. From matlab I would get this message after following the instructions:
vl_compilenn: CUDA: MEX config file: '.../matconvnet/matlab/src/config/mex_CUDA_glnxa64.sh'
gcc-4.4: No such file or directory
However, after compiling from the bash shell the tests would all pass.
This was my command from the command
make ENABLE_GPU=y ARCH=glnxa64 MATLABROOT=/usr/local/MATLAB/R2014a/ CUDAROOT=/usr/local/cuda-5.5/
Monday, January 12, 2015
Friday, October 10, 2014
tune2fs: Bad magic number in super-block while trying to open ... Couldn't find valid filesystem superblock
I was getting this error because I was using it on a exFat ssd hard drive which is used for windows and linux compatibility. If you want to change the label use instead exfatlabel. for instance
sudo exfatlabel /dev/sdb1 ssd
(If exfatlabel is not installed on your system you just need to do:
sudo apt-get install exfat-util)
you can also use sudo fdsik -l to see your hard drives and know which one you need to rename. And use sudo blkid /dev/sdb1 for instance to know what kind of file system you have if you have ntfs you might need to do for instance.
sudo ntfslabel /dev/sdb2 hd1
sudo exfatlabel /dev/sdb1 ssd
(If exfatlabel is not installed on your system you just need to do:
sudo apt-get install exfat-util)
you can also use sudo fdsik -l to see your hard drives and know which one you need to rename. And use sudo blkid /dev/sdb1 for instance to know what kind of file system you have if you have ntfs you might need to do for instance.
sudo ntfslabel /dev/sdb2 hd1
Tuesday, September 16, 2014
unable to open caffe_rcnn_imagenet_model or rcnn_imagenet_deploy.prototxt
If you are getting the above errors you need to change the arguments to
--pretrained_model=../examples/imagenet/caffe_rcnn_imagenet_model
--model_def=../examples/imagenet/rcnn_imagenet_deploy.prototxt
--pretrained_model=../examples/imagenet/caffe_rcnn_imagenet_model
--model_def=../examples/imagenet/rcnn_imagenet_deploy.prototxt
Monday, September 15, 2014
./include/caffe/data_layers.hpp:11:18: fatal error: lmdb.h: No such file or directory compilation terminated.
If you get this error when you run make all for caffe. It is possible that you are missing dependencies. To install missing lmdb dependencies
git clone git://gitorious.org/mdb/mdb.git
cd mdb/libraries/liblmdb
make
sudo make install
./include/caffe/common.hpp:5:27: fatal error: gflags/gflags.h: No such file or directory
If you get this error when you run make all for caffe. It is possible that you are missing dependencies. To install missing gflag dependencies
wget https://github.com/schuhschuh/gflags/archive/master.zip
unzip master.zip
cd gflags-master
mkdir build && cd build
export CXXFLAGS="-fPIC" && cmake .. && make VERBOSE=1
make
sudo make install
Thursday, July 24, 2014
squeeze() takes no keyword arguments on caffe
If you get this error when you run any of the python modules in caffe. It is possible that you have an older version of numpy. You just need to upgrade to the latest version. You can do
sudo pip install -U numpy
Good luck
sudo pip install -U numpy
Good luck
Tuesday, July 15, 2014
mkl_vsl.h: No such file or directory with caffe
I encounter this problem when trying to compile the implementation of NIN on the caffe branch cccp-dev.
To fix the error I had to include a MKL_DIR variable in the Makefile.config file and set it to
MKL_DIR := /opt/intel/mkl
I also included the mkl directory on my LD_LIBRARY_PATH but that didn't seem to help.
After doing that it compiled the code
To fix the error I had to include a MKL_DIR variable in the Makefile.config file and set it to
MKL_DIR := /opt/intel/mkl
I also included the mkl directory on my LD_LIBRARY_PATH but that didn't seem to help.
After doing that it compiled the code
Subscribe to:
Posts (Atom)