Let’s talk about tools, and which one is better than AWS.

AWS has a service called Sagemaker which in this article I am going to explain what it is and what it does.

What is AWS Sagemaker?

As you probably know, machine learning has a cycle and from data to the machine, you need to few steps to work with it.

The first step is to find data and make them understandable and if it’s needed filter them.

Data is being gathered from different sources and they are being stored in databases. If you want to make meaning of them, you need to filter them, sort them and prepare them for your models.

The second step is to use a model, if you read my previous articles, you know in most of the times we are using existing models to train our machine. You just need to make sure you are following the path of the model and giving in it the right structured data.

The third step is to deploy the model to a computer or somewhere in the cloud. This is a part that AWS Sagemaker can be useful and you can use it for your deployments.

What is AWS Sagemaker features?

Basically, it gives you a console of AWS to work with all the things you want by just clicking on buttons so it will be so performant and easy to use.

Sagemaker provides 11 algorithms that you can use to train your models. They are supervised or unsupervised learning use cases. But you are not limited to just these. you can add your own Tensorflow or Apache MXNet Python.

The first step is to create a notebook instance:

and then you’ll go create an S3 bucket because SageMaker reads all the data necessary to train the model from a bucket.

So the next step is to download your data from the S3 bucket, using commands in Jupyter panel.

Then start writing your models and train your machine. It takes a few minutes to train the model and the output goes to the S3 bucket that you specified.

After then you can use your model to test new data and estimate how accurate was your training model.

It’s easy to use and you can find lots of examples here.

If you want to start with basic MINST problem, have a look at these steps which is really straight forward.

Sources:

• https://www.forbes.com/sites/janakirammsv/2019/01/27/amazon-open-sources-sagemaker-neo-to-run-machine-learning-models-at-the-edge/#7a82c96a4d03
• https://aws.amazon.com/sagemaker/features/
• https://aws.amazon.com/sagemaker/