search

wagenaartje / neataptic

:rocket: Blazing fast neuro-evolution & backpropagation for the browser and Node.js
https://wagenaartje.github.io/neataptic/
Other
1.19k stars 279 forks source link

May be a dumb question #119

Open JasonMatthewsDev opened 6 years ago

JasonMatthewsDev commented 6 years ago

This may be a dumb question but after setting up LTSM something like:

let network = new neataptic.architect.LSTM(7, 7, 1)

and I train it with a set of data, what's the effect of training it again with a new set of data? Does it preserve the current network and refine itself with the new training or does it start it over?

alexus2005 commented 6 years ago

It depends if you train the network variable, or create a new network under different alias. If you reuse name "network" for your variable and train it again, it will adjust the existing network.