node-red-contrib-brain 0.1.4

Neural networks in Node-RED

npm install node-red-contrib-brain

node-red-contrib-brain

Install

$ cd $HOME/.node-red
$ npm install node-red-contrib-brain

Or globally alongside Node-RED:

sudo npm install -g node-red-contrib-brain

You will need to restart Node-RED for it to pick-up the new nodes.

Usage

node-red-contrib-brain is based on brain. You need to see it basic usage at first.

Train

When there is no msg.netJSON input, brain node will train with msg.trainData and output net.toJSON in msg.net.

Run

When there is msg.netJSON input, brain node will not train anymore. Instead, it will directory run net.run() with your msg.runData. And output the result in msg.payload.

Options

You could pass a msg.neuralNetworkOptions to brain node. See brain options doc.

Example

Train:

[{"id":"d4a484b.f2b5b78","type":"debug","z":"46d50fa1.b92af","name":"","active":true,"console":"false","complete":"net","x":679,"y":175,"wires":[]},{"id":"3b4d60b3.c4b2a","type":"inject","z":"46d50fa1.b92af","name":"","topic":"","payload":"","payloadType":"date","repeat":"","crontab":"","once":false,"x":150,"y":144,"wires":[["6d766489.92899c"]]},{"id":"6d766489.92899c","type":"function","z":"46d50fa1.b92af","name":"fake data","func":"// This function return a fake json array\nvar trainData = [{input: { r: 0.03, g: 0.7, b: 0.5 }, output: { black: 1 }},\n           {input: { r: 0.16, g: 0.09, b: 0.2 }, output: { white: 1 }},\n           {input: { r: 0.5, g: 0.5, b: 1.0 }, output: { white: 1 }}]\n\nmsg.trainData = trainData\nreturn msg;","outputs":1,"noerr":0,"x":333,"y":183,"wires":[["2fe2efbc.d01d1"]]},{"id":"2fe2efbc.d01d1","type":"brain","z":"46d50fa1.b92af","name":"brain","learningRate":0.3,"errorThresh":0.005,"iterations":20000,"log":false,"logPeriod":10,"x":516,"y":185,"wires":[["d4a484b.f2b5b78"]]}]

Run:

[{"id":"10d56f1d.ef2a91","type":"debug","z":"46d50fa1.b92af","name":"","active":true,"console":"false","complete":"payload","x":687,"y":346,"wires":[]},{"id":"808a1160.7f75f","type":"inject","z":"46d50fa1.b92af","name":"","topic":"","payload":"","payloadType":"date","repeat":"","crontab":"","once":false,"x":131,"y":325,"wires":[["fbed5f0e.0412a"]]},{"id":"fbed5f0e.0412a","type":"function","z":"46d50fa1.b92af","name":"fake data","func":"// This function return a fake json array\nvar netJSON = {\"layers\":[{\"r\":{},\"g\":{},\"b\":{}},{\"0\":{\"bias\":0.5976173927716023,\"weights\":{\"r\":3.5006895738532835,\"g\":-4.542455700505483,\"b\":0.9988932386815509}},\"1\":{\"bias\":0.6470978455858952,\"weights\":{\"r\":3.6115725201557827,\"g\":-4.875546614413311,\"b\":1.211740813346471}},\"2\":{\"bias\":-0.3559477465736521,\"weights\":{\"r\":1.1063595019849224,\"g\":-1.857026678772011,\"b\":0.14886809335684345}}},{\"black\":{\"bias\":3.3336645409591017,\"weights\":{\"0\":-3.7876606581596914,\"1\":-4.023316483216229,\"2\":-1.0819957068479935}},\"white\":{\"bias\":-3.29149645757729,\"weights\":{\"0\":3.782751737648757,\"1\":4.173873416865656,\"2\":0.7154074171638515}}}],\"outputLookup\":true,\"inputLookup\":true}    \nvar runData = { r: 1, g: 0.4, b: 0 }\n\nmsg.runData = runData\nmsg.netJSON = netJSON\nreturn msg;","outputs":1,"noerr":0,"x":315,"y":336,"wires":[["3132e68b.cecd1a"]]},{"id":"3132e68b.cecd1a","type":"brain","z":"46d50fa1.b92af","name":"brain","learningRate":0.3,"errorThresh":0.005,"iterations":20000,"log":false,"logPeriod":10,"x":501,"y":341,"wires":[["10d56f1d.ef2a91"]]}]

License

MIT License