roachsinai

A Tour of PyTorch Internals (Part I)

The fundamental unit in PyTorch is the Tensor. This post will serve as an overview for how we implement Tensors in PyTorch, such that the user can interact with it from the Python shell. In particular, we want to answer four main questions:

1. How does PyTorch extend the Python interpreter to define a Tensor type that can be manipulated from Python code?
2. How does PyTorch wrap the C libraries that actually define the Tensor's properties and methods?
3. How does PyTorch cwrap work to generate code for Tensor methods?
4. How does PyTorch's build system take all of these components to compile and generate a workable application?

Extending the Python Interpreter

PyTorch defines a new package torch. In this post we will consider the ._C module. This module is known as an "extension module" - a Python module written in C. Such modules allow us to define new built-in object types (e.g. the Tensor) and to call C/C++ functions.

roachsinai

#include <functional>
#include <map>
#include <typeinfo>
#include <iostream>

struct Event {
	virtual ~Event() {}
};
struct TestEvent : Event {
	std::string msg;

roachsinai

Lisp interpreter in 90 lines of C++

I've enjoyed reading Peter Norvig's recent articles on Lisp. He implements a Scheme interpreter in 90 lines of Python in the first, and develops it further in the second.

Just for fun I wondered if I could write one in C++. My goals would be

1. A Lisp interpreter that would complete Peter's Lis.py test cases correctly...
2. ...in no more than 90 lines of C++.

Although I've been thinking about this for a few weeks, as I write this I have not written a line of the code. I'm pretty sure I will achieve 1, and 2 will be... a piece of cake!