Neural networks. Why Python?

Question

Python

Good day. In almost all questions about the language for neural networks, the only answer is python. But why? Why, among a bunch of languages, it was he who became it?

Answer 1

You mean about the situation when neural networks are all research projects.
Scientists are not necessarily programmers. And Python is a programming language with a low barrier to entry.
Nobody writes from scratch, does not waste time - they just use libraries already written by others.
As a result, a neural network community has developed around Python.
If you just for research purposes - do the same.
PS:
You still don't know about Fortran.
The language has not been used by programmers for a long time.
But physicists continue to be used.

Answer 2

Because in addition to simplicity, conciseness and expressiveness, which allow developing complex algorithms with minimal time and effort, the language also has a powerful interoperability mechanism with C \ C ++, which opens up access to fast calculations.

Answer 3

Because neural networks are, as a rule, programs that are not very large in volume, but which often have to be changed in order to select the optimal architecture, data preprocessing, and so on. Therefore, there are almost no problems with legacy code, but there are requirements for rapid development. Python does this better than C++ or Java.
Machine learning frameworks are often a Python or Torch frontend and a C++ backend.

Answer 4

Python - a quick understanding of the subject area and nothing more!
Keras.io (+ BONUS: how to install )
(translation)
Read the Keras.io documentation.
Keras is compatible with: Python 2.7-3.6.
Want to use GPU speed NORMALLY - C/C++ and NVIDIA CUDA Toolkit only .