Preparing the platform

While data input has a big impact on model quality, the hardware/software platform where we train/validate/test the model will also impact the model and the development process. Choosing the right platform is very important to the ML process.

While certainly, you can choose to use a desktop or laptop for ML model development, it is a recommended practice to use cloud platforms, thanks to the great advantages that cloud computing provides: self-provisioning, on-demand, resilience, and scalability, at a global scale. Many tools are provided in cloud computing to assist data scientists in data preparation and model development.

Among the cloud service providers, Google Cloud Platform provides great ML platforms to data scientists: flexible, resilient, and performant, from end to end. We will discuss more details of the Google Cloud ML platform in the third part of the book.

Now that we have prepared the datasets and ML platform, let’s dive right...