About

The repository is created to support blogpost on Custom Sagemakers. There are three steps needed to deploy the solution in AWS.

1. Train&deploy custom model with SageMaker
2. Create API that would use SageMaker model
3. Build and deploy a Flask application which would take the tweets and return results of API

Step 1 is covered in the blogpost. Steps 2 and 3 available in corresponding tutorials. During step 3 it is possible to use the code of the Flask application which is stored in the folder EC2_Flask_app/.

Parameters to set up

Please set the following parameters:

1. 'PATH_TO_DATA' in file SageMaker/mysrc/serve.py
2. url in file EC2_Flask_app/app.py