From the course: Learning Amazon SageMaker
Unlock the full course today
Join today to access over 22,400 courses taught by industry experts or purchase this course individually.
Model training
- [Instructor] Now that we've prepared the model data sets, the two CSV files, and uploaded them to S3, we'll look to start creating the training instance and start actually training the model. The way that we do that is initially using the SageMaker package, we import the get_image_uri function, and we use that to (mumbles) using boto 3, getting the XGBoost container. This container will contain all the code that's required for the XGBoost model to start training on a separate instance. Because we've already created the train and validation CSV data sets, we'll create objects that reference those using the sagemaker.s3_input function, which we've done here. The next cell arguably contains the bulk of the training functions. These next four functions will create a SageMaker session, will create an estimator object, taking the container that we created before, the role that we created initially, and all the other values to actually create an XGB object which will allow us to start…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.