Training and saving a model

Engineering NuPIC
1

Hello everyone,
I am trying to train CLA model on 1000 records and save the trained model. Can anyone share the example source code of training and saving a model. I want to do the prediction based on the saved model. I have thought of enable learning for the 30% of the data and do the prediction on remaining 70% if the data is not streaming.

Thanks in anticipation.

1 Like
2

Here is a simple example of feeding data from a CSV file into a model.

There are no training vs testing phases for HTM because it’s an online learning system. I suggest you leave learning on all the time for prediction. That way it can continue to adjust to the live data as it learns.

For details on how to save/load models, see our FAQ.

1 Like
3

Hi @rhyolight ,
Thanks for the reply.

I have gone through that example but i am not able to understand that where are we doing the training there.
Does the model get updated when we call model.run(inputData) . Where should i use the model.save(DirectoryPath) in order to save the trained model and doing the offline prediction. Please clarify me.

4

Yes, it learns on each iteration unless you call disableLearning().

You can save anytime.

I don’t understand the question. You do not have to save a model unless you want to shut down the program and start it up again later with the same model state. Saving / loading a model does not affect its learning state.

5

@rhyolight Thanks again.

My issue got resolved. I was having problem in saving the trained model. i am able to save the model now.