topicsimilaritymodel_demo

Application that demonstrates the Topic Similarity Model in action.

1. Retrieve a small set of sample text documents and "train" the model.
2. Retrieve the text of what is to be matched (aka, the "topic" text).
3. Retrieve the larger set of unknown documents and insert them into the model for matching.
4. Submit a job to the Gravity-AI container and evaluate the results.

Prerequisites

• Docker
• A running instance of the Gravity AI container with the Topic Similarity Model

Installation

docker-compose up

Execution

This demo runs on an instance of Jupyter Notebook. To access the notebook, look for the URL on the console that launched the Docker environment. It should appear as something like this:

jupyter       |     To access the notebook, open this file in a browser:
jupyter       |         file:///home/jovyan/.local/share/jupyter/runtime/nbserver-16-open.html
jupyter       |     Or copy and paste one of these URLs:
jupyter       |         http://0b90cc578f8a:8888/?token=e04fb77ae092f52dd1a31bdc24d03f79933d0db1f620e77c
jupyter       |      or http://127.0.0.1:8888/?token=e04fb77ae092f52dd1a31bdc24d03f79933d0db1f620e77c

Launch the URL from a browser and navigate to the Gravity_AI_Container_Demo_MoviePlots.ipynb file.

Run the steps in the file, via the Jupyter Notebook UI.