TensorFlow + Node.js

TensorFlow is Google's machine learning and data flow-based computation system. It is implemented as C++ runtime, and comes with Python bindings out-of-the-box, including a comprehensive toolbox of primitives for use in defining a variety of models, including various kinds of neural networks.

This module makes it possible to tap into those capabilities from node.js, with a 100% JavaScript (or TypeScript if you prefer) developer experience. It provides a natural node.js development experience, while taking care of the interop.

Essentially it allows you to define graphs, work with tensors and operations, and then execute that graph to run through your data within a session. These key concepts define the essence of the programming model.

A glimpse at what is in the works

This is project is super early, so the current implementation is far from complete, and has been cobbled together somewhat quickly to share early ideas, and gather community input and participation.

But, anyway, here is what is in the works:

var tf = require('tensorflow'),
    fs = require('fs');

// Define the graph
var g = new tf.Graph();
var shape = [2, 2]
var p1 = g.placeholder(tf.types.float, shape).named('p1');
var p2 = g.placeholder(tf.types.float, shape).named('p2');
var value = g.matmul(p1, p2).named('value');

// Optionally save it out (with corresponding APIs to load, instead
// of re-building the graph, for example when using the resulting model).
fs.writeFileSync('/tmp/hello.graph', g.save());

// Execute the graph
var session = new tf.Session(g);

var data = {};
data[p1] = new tf.Tensor([[1.0, 0.0],[0.0, 1.0]]);
data[p2] = new tf.Tensor([[3.0, 3.0],[3.0, 3.0]]);

var results = session.run([ value ], data);
console.log(results[value]);
```javascript


As it starts to come together it will be available on
[npmjs](https://www.npmjs.org/package/tensorflow) as any other module.

## More for later, and helping ...

Stay tuned for more. If you're interested in helping out, by all means, 
please connect here. I'll share a bit more detail about the roadmap, as well
as issues for big questions, design issues and areas of exploration.