Training with code written in popular frameworks
SageMaker's fully managed training works with your favorite ML frameworks too, thanks to the container technology we mentioned previously. You may have been working with Tensorflow
, PyTorch
, Hugging
Face
, MXNet
, scikit-learn
, and many more. You can easily use them with SageMaker so that you can use its fully managed training capabilities and benefit from the ease of provisioning right-sized compute infrastructure. SageMaker enables you to use your own training scripts for custom models and run them on prebuilt containers for popular frameworks. This is known as Script Mode. For frameworks not covered by the prebuilt containers, you also can use your own container for virtually any framework of your choice.
Let's look at training a sentiment analysis model written in TensorFlow as an example to show you how to use your own script in SageMaker to run with SageMaker's prebuilt TensorFlow container. Then we will describe...