danchern97 / intent_detection

Code provided for Bachelor diploma "User intent detection in conversation systems", 2019, MIPT.

User intent detection in conversation systems.

This is the code provided for Bachelor diploma "User intent detection in conversation systems", 2019, MIPT.

Deployment

First build a docker container: docker build -t intent_catcher .

Then run it: docker run -it -p 8014:8014 intent_catcher

To send phrases to container, use python3 test_server.py. To modify phrases, see test_server.py and modify the request sentences.

Intent data is located at data/full_dataset.json at data['alexa_prize']['train']. The list of intents is located there.