Simple code and model for making sentence-level speech-act predictions (e.g. statement, question, etc.) Initial model is a convolutional architecture trained on the SWDA conversation corpus.
Build docker image by docker-compose in ./docker_clf_project dir: docker-compose build
Run the container using: docker run -it -p 50051:50051 <docker image id>
By default, gRPC connection is via port 50051
Test the gRPC server. In a separate command window run the script python src/spam_clf_client.py
This sends two separate messages to the server. You should see a set of predicted speech-acts for
each set of messages.