In this section, we will use the Amazon SageMaker image classification algorithm to train on the cifar10 dataset.
The Amazon SageMaker image classification algorithm is a supervised learning algorithm that supports multi-label classification. It takes an image as input and outputs one or more labels assigned to that image. It uses a convolutional neural network (ResNet) that can be trained from scratch or trained using transfer learning when a large number of training images are not available.
To train a model in Amazon SageMaker, you can use either the
Amazon SageMaker Python SDK or the
AWS SDK for Python (Boto 3). In the last two labs, we used
AWS SDK for Python (Boto 3) to train the model. In this lab, you will use the Boto 3 to train the model. Amazon SageMaker uses the
CreateTrainingJob API to run the training.
MLAI/built-in-algorithm, double click on
cifar10.ipynbto open it.
Please notice that you can still see the progress of your training job and the metrics, even though you have not setup the experiment in the notebook when you built the estimator. You can click on the
Unassigned trial component section of the
SageMaker Experiment List tab in the left sidebar.
Then, you can right click on the training job and either deploy it with one click or check the details of the training.
You can see the details of training such as charts (e.g. validation accuracy per epoch), metrics, parameters, etc.