You can write fast code in C++. You can write code fast in Python.
Python has emerged as the de facto programming language engineering, data-science and machine-learning disciplines. Increasingly IoT appliances use Python too, for high-level control. Python is nominally ten times slower than C or C++, so how is it that these solutions often exhibit great performance?
The secret is that many core data-science, machine learning, computer vision, and computational geometry libraries are implemented in C++, and then later exposed to Python.
In this talk, we introduce pybind11 a header-only, open-source C++11 library for exposing existing C++ libraries as Python modules. Come along for plenty of code examples, and advice on building hybrid Python and C++ solutions.