2 ms·
Just one reason is that some Python libraries (numpy, tensorflow, pytorch) allow you to work with high dimensional arrays (3-4 dimensions) without for loops. M
by WithinReason 5y ago
Just one reason is that some Python libraries (numpy, tensorflow, pytorch) allow you to work with high dimensional arrays (3-4 dimensions) without for loops.
ML also needs reverse autodifferentiation, which would be a real pain in C++.