3 ms·
Pretty much this. Add to that pybind11's exceptional numpy compatibility (including easily mapping C types to structured numpy dtypes that can be e.g. converted
by aldanor 7y ago
Pretty much this. Add to that pybind11's exceptional numpy compatibility (including easily mapping C types to structured numpy dtypes that can be e.g. converted to a dataframe), and there's not many choices left if you're not planning on reinventing a wheel or two.
(disclaimer: pybind11 contributor)
- pletnes 7y agoIs pybind numpy aware? This is interesting. Having never used pybind, where can I learn more?
- Nimitz14 7y agoCheck the github and the docs. Funny reading this now, I literally just implemented something in 15min using C++/pybind11 (and operates on numpy arrays) that is an order of magnitude faster than when I used numpy operations.