Saving and loading models

by: Kevin Broløs & Chris Cave
(Feyn version 2.1 or newer)

Models can be saved to a file for use at a later time or place.

Note: Due to incompatibilities in the different versions of the QLattice in version 2.1, you are unable to load any models saved prior to version 2.1. If you need access to older models, we recommend you use a previous version of feyn.

Here is an example how:

import feyn
from feyn.datasets import make_classification

# Generate a dataset and put it into a dataframe
train, test = make_classification()

# Connect to a QLattice and run a classification simulation
ql = feyn.connect_qlattice()
models = ql.auto_run(
    data=train,
    output_name='y',
    kind='classification'
)

# Save a model to a file
models[0].save('my_model.json')

You can load the Model using feyn.Model.load.

from feyn import Model

model = Model.load('my_model.json')

prediction = model.predict(test)

Once a Model is saved and selected, you can load them into any Python environment to do predictions. You don't need access to a QLattice for this.

A loaded Model is no different from a Model sampled from the QLattice. For example, you can resume fitting a Model after loading it.