In this project I'm going to explain how was my first (and second, and third ++) contact with IRIS IntegratedML and use IRIS to compose my Machine Learning Environment.
I decided to take the the suggestion on this post: https://community.intersystems.com/post/posts-and-tags-problem-intersystems-iris-ai-contest
Actually two problems.
I have deployed the trained model as a demo here: http://iris-ml-suite.eastus.cloudapp.azure.com/
I start exploring the data to extract some features from the Post dataset. To accelerate I had used the Tags field as training set.
Type or paste a community post on the textarea of site and click in Predict. After this, my model will predict which tags should fit better for it.
Clone my repository typing these commands:
git clone https://github.com/renatobanzai/iris-ml-suite.git
adjust docker settings to up memory and cpu the AI demands more capacity
docker-compose build docker-compose up -d
1st time running will depend of your internet link to download the images and dependencies. If it last more than 15 minutes probably something goes wrong feel free to communicate here. After the 1st time running the next ones will perform better and take less then 5 minutes.
After a while you can open your browser and go to the address:
I'm using for now the PYTHON namespace
user: _SYSTEM pass: SYS