# On Ubuntu 14.04, if cuda tools are installed via # CUDA directory contains bin/ and lib/ directories that we need. the default for Linux is g++ and the default for OSX is clang++ # To customize your choice of compiler, uncomment and set the following. # possibility of simultaneous read and write # You should not set this flag if you will be reading LMDBs with any # uncomment to allow MDB_NOLOCK when reading LMDB files (only if necessary) # uncomment to disable IO dependencies and corresponding data layers
# CPU-only switch (uncomment to build without GPU support). # cuDNN acceleration switch (uncomment to build with cuDNN). # Contributions simplifying and improving our build system are welcome! Since I do not have any eGPU right now, I want to use just the CPU. This is my nfig file, similar to this other one I found on the Internet. Ld: library not found for -lboost_thread-mtĬlang: error: linker command failed with exit code 1 (use -v to see invocation) usr/local/include/glog/logging.h:943:30: note: expanded from macro 'LOG_EVERY_N'Īt the end of the process, it reported the following error:Ĭlang: warning: argument unused during compilation: '-pthread' LOG_EVERY_N(WARNING, 1000) << "DEPRECATED: ForwardPrefilled() " include/caffe/net.hpp:41:5: warning: unused typedef 'INVALID_REQUESTED_LOG_SEVERITY' In file included from tools/extract_features.cpp:9: The terminal threw multiple warnings such as this down below: The problem came when typing make all -j8. Although I had already installed those two packages within the Anaconda environment.
using Brew for general dependecies, including protobuf and boost. I followed the official instructions to install Caffe–i.e. I am running Python 3.6 via Anaconda 4.3.23 on Mac OS Sierra 10.12.6.
0 kommentar(er)
