Amazon SageMaker

Amazon SageMaker

Input

• Supported content type: text/plain.
• The input file has to be in utf-8 encoding only
• The algorithm works with any English text data with a word limit in range 100 to 250 words.

Output

• Content type: application/zip.
• A zipped folder contains two output files, a “.csv” file with the Triplets and a “.png” file with the knowledge graph.
• The csv will have the triplets {Subject-Predicate-Object} and a graphical representation of these triplets.
• In the knowledge graph (.png), the nodes represent Subject & Object and edges represent Predicate.
• The nodes in the knowledge graph is color coded based on NER tags:
  • Location: Green
  • Org: Maroon
  • Date: Red
  • Person: Blue
• Other NER tags:
  • Skyblue
• Non NER tags:
  • Yellow

Invoking endpoint

AWS CLI Command If you are using real time inferencing, please create the endpoint first and then use the following command to invoke it:

aws sagemaker-runtime invoke-endpoint --endpoint-name "endpoint-name" --body fileb://Input.txt --content-type text/plain --accept application/zip output.zip

Substitute the following parameters:

• "endpoint-name" - name of the inference endpoint where the model is deployed.
• Input.txt - Input file.
• text/plain - MIME type of the given input file.
• output.zip - filename where the inference results are written to.

Resources

• Sample Input
• Sample Jupyter Notebook
• Sample Output

Mphasis DeepInsights
Sample Jupyter Notebook
Sample Input File