Closed ostjul closed 5 months ago
@ostjul I had the same problem but was able to fix it by copying the include
dir from my build folder after the fvdb install. See below for what I ran.
cd openvdb/fvdb
cp -R build/lib.linux-x86_64-cpython-39/fvdb/include fvdb/include
Thanks a lot @rfeinman, it worked after copying all files from fvdb/include/nanovdb/nanovdb
to fvdb/include/nanovdb
Hi, this coud also be fix by:
# Install sparse 3D deep learning framework
git clone https://github.com/AcademySoftwareFoundation/openvdb.git
cd openvdb
git fetch origin pull/1808/head:feature/fvdb
git checkout feature/fvdb
rm fvdb/setup.py && cp ../assets/setup.py fvdb/
cd fvdb && python setup.py develop
cd ../..
# Mesh extraction
cd ext/nksr-cuda
python setup.py develop
cd ../..
@xrenaa Thanks for your guidelines! I followed the provided instructions and also copied the files as @rfeinman has mentioned before, but I met a problem:
x86_64-conda-linux-gnu-c++: error: unrecognized command-line option '-march'
It seems that the c++ version is not suitable, how can I fix this bug?
Thanks for your reply!
Hello, after following the above steps and running the infer code, I encountered the following error:
openvdb/fvdb/fvdb/libfvdb.so: undefined symbol: _ZN2at4_ops13scalar_tensor4callERKN3c106ScalarESt8optionalINS2_10ScalarTypeEES6_INS2_6LayoutEES6_INS2_6DeviceEES6_IbE
Can you please help me solve this issue?
@ostjul I had the same problem but was able to fix it by copying the
include
dir from my build folder after the fvdb install. See below for what I ran.cd openvdb/fvdb cp -R build/lib.linux-x86_64-cpython-39/fvdb/include fvdb/include
Could you please tell me how to install fvdb? I using pip install fvdb failed.Thanks
@ostjul I had the same problem but was able to fix it by copying the
include
dir from my build folder after the fvdb install. See below for what I ran.cd openvdb/fvdb cp -R build/lib.linux-x86_64-cpython-39/fvdb/include fvdb/include `
Could you please tell me how to install fvdb? I using pip install fvdb failed.Thanks
Solved according to this:
Hi, this coud also be fix by:
# Install sparse 3D deep learning framework git clone https://github.com/AcademySoftwareFoundation/openvdb.git cd openvdb git fetch origin pull/1808/head:feature/fvdb git checkout feature/fvdb rm fvdb/setup.py && cp ../assets/setup.py openvdb/fvdb/ cd fvdb && python setup.py develop cd ../.. # Mesh extraction cd ext/nksr-cuda python setup.py develop cd ../..
直接在 xcube 的根目录下执行,不过要把一行进行修改:
rm fvdb/setup.py && cp ../assets/setup.py openvdb/fvdb/
to
rm fvdb/setup.py && cp ../assets/setup.py fvdb/
Thanks for the code release and great paper! The setup of fvdb seems to work fine on my system with one exception, that I could fix. Following your installation instructions on a linux server I experienced the following build error:
cannot find external/c-blosc/install/lib/libblosc.a: No such file or directory
A solution that worked for me was changing the path in the setup file. In
openvdb/fvdb/setup.py
line 264:(...) extra_objects=['external/c-blosc/install/lib/libblosc.a'] + cudnn_static_libs,
needs to be changed toextra_objects=['external/c-blosc/install/lib64/libblosc.a'] + cudnn_static_libs,
if the systems default is building the 64bit version in lib64. Otherwise, it fails to findlibblosc.a
. Seems to be related to an easybuild/GCC compiler issue.While I can import fvdb python packages now in the nksr-cuda setup
#include <fvdb/GridBatch.h>
doesn't work and I could not figure out the right referencing myself yet.