Web application to demonstrate ability of some state-of-the-art machine learning models at the questions answering task. The models chosen for the first deployment of the app are:

1. BiDAF - ONNX pre-trained model
2. DistilBERT - HuggingFace implementation
3. RoBERTa - Deepset implementation (on Hugging face model zoo)
4. ALBERT - twmkn9 implmentation (on Hugging face model zoo)

The demo deployment will likely utilize Google Build to containerize the application, Google Container Registry for storing and managing a container and Google Cloud Run to deploy it as a web endpoint.

More about Google Cloud Run

You need Python 3 to run this application. Other dependencies can be found in the requirements.txt file.

Machine Comprehension uses the MIT license.