Intro to TensorFlow

Cloud Machine Learning Engine

• Train, deploy, and productionalize ML models at scale

Estimators

• Interchangeable
• Test many standard pre-made estimator models in quick succession

Checkpoints

Feature Engineering

Scale to large datasets

Find good features

Preprocessing

Art and Science of Machine Learning

Parameter Norm Penalties

• L1 / L2 regularization
• Max-norm regularization

Hyperparameter Tuning

• ML models - Functions with parameters and hyper-parameters
• Sensitive to batch size and learning rate (dataset-dependent)

Logistic Regression

• Use Cross Entropy typically