Keras & tf.Keras

AimCallback for Keras is designed to enhance your experiment logging and monitoring. It thoroughly records essential information, including hyperparameters, training, validation, and test time metrics like loss and accuracy. Moreover, it offers comprehensive system usage tracking, keeping an eye on CPU and GPU memory utilization.

This tutorial leverages the well-known handwritten digit recognition task to describe how to integrate Aim with Keras & tf.Keras to train a digital image classification model based on the mnist dataset.

It only takes 2 steps to easily integrate aim in Keras to record experimental information.

from aimstack.experiment_tracker.keras import Callback as AimCallback

In Keras, we call the fit() method of the model object to train the data. The callbacks are provided here. AimCallback inherits the usage specification of callbacks. We just need to add it to the callbacks list.

model.fit(x_train, y_train, epochs=5, callbacks=[
          # in case of tf.keras, we use aim.tensorflow.AimCallback
          AimCallback(experiment="test_experiment")
])

See AimCallback source here. Check out a simple example using Keras here. Check out a simple example using tf.Keras here.