Basic Virtual Assistant

Basic Virtual Assistant is the accompanying Github repository for my blog series "Making Your Own Alexa". The code in this repository can be used as a basis for a virtual assistant project.

Part 1: https://towardsdatascience.com/making-your-own-alexa-entity-extraction-8c7f23eb65a

Getting Started

pip install -r requirements.txt

NLTK will require you to download a data for its tokenizer so open Python in your terminal and run the following code

import nltk
nltk.download('punkt')

Usage

The intent_trainer module contains a CLI for training new intents. It has a few conditional options that can be used:

• task is the task you want to run (train or predict).
• --schema_file is the path to the training data you want to use. E.g. commands/play_commands.json.
• --name is the name of the intent that you're training.
• --sample_size is the amount of rows that should be saved so that they can used to train the intent classification model.
• --batch_size is the size of the mini-batches.
• --epochs is the number of epochs the model should train for.
• --command is the command you want to predict the labels for.

Below is an example on how you can train an Named Entity Recognition model for the Play command.

python intent_trainer.py train --schema_file commands/play_commands.json --name play --sample_size 400 --batch_size 128 --epochs 15

And here is an example on how to test one of your commands.

python intent_trainer.py predict --name play --command "play let it be by the beatles"