Training models with SageMaker's built-in algorithms
When you want to build an ML model from a notebook in SageMaker Studio for your ML use case and data, one of the easiest approaches is to use one of SageMaker's built-in algorithms. There are two advantages of using built-in algorithms:
- The built-in algorithms do not require you to write any sophisticated ML code. You only need to provide your data, make sure the data format matches the algorithms' requirements, and specify the hyperparameters and compute resources.
- The built-in algorithms are optimized for AWS compute infrastructure and are scalable out of the box. It is easy to perform distributed training across multiple compute instances and/or enable GPU support to speed up training time.
SageMaker's built-in algorithm suite offers algorithms that are suitable for the most common ML use cases. There are algorithms for the following categories: supervised learning, unsupervised learning...
