Step 6: Deploy Your Model

You can train, tune and deploy machine learning models on Google Cloud. See the following resources for guidance on deploying your model to production:

  • Tutorial on how to export a Keras model with TensorFlow serving.
  • TensorFlow serving documentation.
  • Guide to training and deploying your model on Google Cloud.

Please keep in mind the following key things when deploying your model:

  • Make sure your production data follows the same distribution as your training and evaluation data.
  • Regularly re-evaluate by collecting more training data.
  • If your data distribution changes, retrain your model.