search

liquidcarrot / carrot

🥕 Evolutionary Neural Networks in JavaScript
https://liquidcarrot.io/carrot/
MIT License
295 stars 34 forks source link

Add a data layer #54

Open luiscarbonell opened 5 years ago

luiscarbonell commented 5 years ago

Is your feature request related to a problem? Please describe.

Yes, I have to keep dealing with the importing of data and transforming thereof using a bunch of other libraries - when trying to train on historic data. It's a lot of coding that keeps getting repeated.

Describe the solution you'd like

A built in data import system that auto-magically imports a .csv file and splits the data into a test and train chunk, depending on the size of the file(s); and auto-magically creates the network around the data-set.

Describe alternatives you've considered

I've been using fast-csv to consume the .csv files, lodash for all the text transformations.

Additional context

I've been using a lot of the data sets provided in this article.

dan-ryan commented 5 years ago

Tensorflor.js has a helpful CSV method, so that would be great if you can add something similar. https://js.tensorflow.org/api/latest/#data.csv

luiscarbonell commented 5 years ago

@dan-ryan Thank you! That is a solid reference.

I ran into this repo, mappr - and it is inspired me to think about being able to map directly from a MondoDB Schema into Carrot.

That might make the process a lot easier....just letting the idea float around for now, while I test it on .csv files.