search

AaronJackson / vrn

:man: Code for "Large Pose 3D Face Reconstruction from a Single Image via Direct Volumetric CNN Regression"
http://aaronsplace.co.uk/papers/jackson2017recon/
MIT License
4.52k stars 746 forks source link

./run.sh error #112

Closed fpradom closed 5 years ago

fpradom commented 5 years ago

After installing the package in Fedora 29 I have following error on running run.sh: $ ./run.sh ImportError: numpy.core.multiarray failed to import /home/f_prado/usr/local/torch/install/bin/luajit: ...ado/usr/local/torch/install/share/lua/5.1/trepl/init.lua:389: ...ado/usr/local/torch/install/share/lua/5.1/trepl/init.lua:389: ...ado/usr/local/torch/install/share/lua/5.1/trepl/init.lua:389: Python error: import numpy array module ImportError: numpy.core.multiarray failed to import

stack traceback: [C]: in function 'error' ...ado/usr/local/torch/install/share/lua/5.1/trepl/init.lua:389: in function 'require' main.lua:8: in main chunk [C]: in function 'dofile' ...ocal/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk [C]: at 0x00406c80 ls: no se puede acceder a '.txt': No such file or directory THCudaCheck FAIL file=/home/f_prado/usr/local/torch/extra/cutorch/lib/THC/THCGeneral.c line=70 error=30 : unknown error /home/f_prado/usr/local/torch/install/bin/luajit: ...ado/usr/local/torch/install/share/lua/5.1/trepl/init.lua:389: ...ado/usr/local/torch/install/share/lua/5.1/trepl/init.lua:389: cuda runtime error (30) : unknown error at /home/f_prado/usr/local/torch/extra/cutorch/lib/THC/THCGeneral.c:70 stack traceback: [C]: in function 'error' ...ado/usr/local/torch/install/share/lua/5.1/trepl/init.lua:389: in function 'require' process.lua:16: in main chunk [C]: in function 'dofile' ...ocal/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk [C]: at 0x00406c80 ls: no se puede acceder a '.raw': No such file or directory [f_prado@localhost vrn]$ python Python 2.7.15 (default, Oct 15 2018, 15:26:09) [GCC 8.2.1 20180801 (Red Hat 8.2.1-2)] on linux2 Type "help", "copyright", "credits" or "license" for more information.

Can you help me with it? Thanks in advance

fpradom commented 5 years ago

vrn package installation details

installed CUDA 9.0, CUDNN 9.0 python 2.7 pip install vivis imageio dlib matplotlib numpy numpy installed version : 1.15.4 ---- MODIFY FLOATN.H ---- edit /usr/include/bits/floatn.h modified adding : 34 # define __HAVE_FLOAT128 0 35 #endif

if CUDART_VERSION

undef __HAVE_FLOAT128

define __HAVE_FLOAT128 0

endif

37 / Defined to 1 if __HAVE_FLOAT128 is 1 and the type is ABI-distinct 38 from the default float, double and long double types in this glibc. / 39 #if __HAVE_FLOAT128

---- INSTALLING TORCH ----

for torch 7 used gcc 5.3 instaed of the generally used gcc 7.3 added : export TORCH_NVCC_FLAGS="-D__CUDA_NO_HALF_OPERATORS__" before ./install.sh

---- Install fb.python ----

edited : /home/f_prado/usr/src/fblualib/fblualib/python/CMakeLists.txt modify line 66 : SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=gnu++11 -fPIC")

AaronJackson commented 5 years ago

Modifying system headers is a terrible idea. Please no one do this.

As the README states, use CUDA 8.0, not 9 or above.

fpradom commented 5 years ago

Thanks, I will try with your suggestions.