Save and Load

Saving and loading your trained model is critical. No one wants to spend hours training to just use there model once. EasyNN handles saving and loading in a few lines.

Save:

Once your model has been trained. This is all you will need to save.

model.save("name")

Saving the model will output two files. The parameters file are all the models optimized numbers. The structure is a pickle file that stores most of the code necessary to run a model.

name_parameters.npz
name_structure.pkl

from EasyNN.examples.cifar10.trained import model
from EasyNN.examples.cifar10.data import dataset

images, labels = dataset

# Show the image
model.show(images[2])

Click here to see a Save full example.

Loading Model:

To load a model its only two lines. Importing the base model class and loading your saved model by using its name.

from EasyNN.model import Model

model = Model.load("name")

Click here to see a Load full example.