reduce deltas to zeros just before weights are set to them

`trainTimeMs` is gone and will be in a new pr

`log` doesn't accept functions but will be in a n...

When my friend wanted me to give him a Charmander tattoo and he got this https://timedotcom.file...

Added transpose and test runner for layer namespace

Simplified log test, and added in the ability to pass in custom log method.

I thought this was ...

Adds back the () => in tests

Ran `npm run make` to get things

Converts the async method to a promise, not a cb

Updates tests

Helps code be more dry

### Makes training async via thaw.js

### Removes `reinforce` option
I felt this cleans up and s...

Fix multiple run issue

Fix for #102 gpu exploding

sp is important

I didn't add anything in the readme so I added a link to where people can create issues.

I think having a template for issues and pull request help keep things focused and productive. ...

some minor cleanup ensuring learningRate is sent to praxis, and that static properties are on class

We were iterating twice, once inside the neural network, and once outside. This resolves it. Al...

fix #116 and bump version number

Simplify unit tests for checking if number comes out of net.

Simplify unit tests for checking if number comes out of net.

Also unit test.
Bump version number.

change data.constructor !== Array to Array.isArray

using search and repalce: ([a-z0-9A-Z\.\[\]]+)\.constructor\s*!==\s*Array to Array.isArray\($1\)

and don't define runInput and calculateDeltas as null if the prototype has them set

Nn gpu

forward propagate is connected

Added link to tutorial