Luis Carbonell

Results 54 comments of Luis Carbonell

Maybe, we could validate a lot of parameters and throw more descriptive errors with something like a JSON schema and using an ES6 parameter schema, mixes with tools like [`schm`](https://github.com/diegohaz/schm)...

Related to #30 Will also add: https://parallel.js.org/ here

@christianechevarria I find it easier - conceptually - to imagine `architect.[architecture]` - as just an extension of `Network`. Would you be cool with - or do you think it's easier...

@tracycollins Out of curiosity what are you using to add the additional "object wrapping" functionality? _I was thinking of adding this using [`mappr`](https://www.npmjs.com/package/mappr) or similar functionality._

@tracycollins I went ahead and created a [project](https://github.com/liquidcarrot/carrot/projects/18?add_cards_query=is%3Aopen) so I can start "knocking it out": https://github.com/liquidcarrot/carrot/projects/18 Do you have an ideal "interface" you would like? **Example** >Maybe something like this....?...

@tracycollins Would the following "work"? **Node Stream** ```javascript const nodestream = new Duplex(); const neat = new Neat(); neat.evolveFromStream(nodestream) .on("iteration", (stuff) => console.log("Ran one generation")) .on("finish", (information) => console.log("Just finished...

>@luiscarbonell How would training end using node stream? close the stream? @tracycollins Yeah, if I remember correctly, Node ends streams by passing `null` - I think the WebStream MDN API...

Just putting this here for near future reference: - https://arxiv.org/abs/1510.00149 - http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.57.4343&rep=rep1&type=pdf - https://datascience.stackexchange.com/questions/26755/cnn-how-does-backpropagation-with-weight-sharing-work-exactly - https://www.quora.com/What-exactly-is-meant-by-shared-weights-in-convolutional-neural-network

@dan-ryan Are you still having this issue? I can't replicate it on my end anymore. Code snippet?

@dan-ryan The tests were running fine on our end now - are you still running into the issue?