Deep Learning From Scratch Using Modern C++

Pre install

Use std::execution::par parallel execution. They are used to specify the execution policy of parallel algorithms. so, install tbb!

foo@bar:~$ sudo apt -y install libtbb-dev

Build

Build with Clang-12.

foo@bar:template/for-class$ bazel build --config=clang_config  //main:hello-world

Test

Test with uftrace and valgrind.

Install uftrace and valgrind in ubuntu.

foo@bar:~$ sudo apt -y install uftrace valgrind

uftrace

Test hello-world program with uftrace in terminal.

foo@bar:template/for-class$ uftrace ./bazel-bin/main/hello-world

valgrind

Test hello-world program with valgrind in terminal.

foo@bar:template/for-class$ valgrind ./bazel-bin/main/hello-world

lint

Install cpplint.

foo@bar:~$ sudo apt -y install python3  && pip install cpplint
foo@bar:~$ echo "export PATH=\$PATH:\$HOME/.local/bin" >> .bashrc && source .bashrc

And do lint.

foo@bar:template/for-class$ cpplint main/hello-world.cc

Reference

『밑바닥부터 시작하는 딥러닝』(한빛미디어, 2017)

Installing Bazel on Ubuntu

Bazel Tutorial: Build a C++ Project

Bazel Tutorial: Configure C++ Toolchains

Bazel command line options