This example is based on Keras-Tuner CIFAR10 sample to demonstrate how to run HP tuning jobs using TensorFlow Cloud and Google Cloud Platform at scale

HP Tuning CIFAR10 on Google Cloud with tensorflow_cloud and CloudTuner

submited by
Style Pass
2021-05-19 13:50:32

This example is based on Keras-Tuner CIFAR10 sample to demonstrate how to run HP tuning jobs using TensorFlow Cloud and Google Cloud Platform at scale.

In this section we define our tuning parameters using Keras Tuner Hyper Parameters and a model-building function. The model-building function takes an argument hp from which you can sample hyperparameters, such as hp.Int('units', min_value=32, max_value=512, step=32) (an integer from a certain range).

In this section we configure the cloud tuner for both remote and local execution. The main difference between the two is the distribution strategy.

This step will prepare your code from this notebook for remote execution and start NUM_JOBS parallel runs remotely to train the model. Once the jobs are submitted you can go to the next step to monitor the jobs progress via Tensorboard.

Most remote training jobs are long running, if you are using Colab it may time out before the training results are available. In that case rerun the following sections to reconnect and configure your Colab instance to access the training results. Run the following sections in order:

Leave a Comment