/Bavardez

Building a chatbot from scratch.

Primary LanguagePythonMIT LicenseMIT

Bavardez

Building a chatbot from scratch. The bot uses a Bag Of Word for encoding (no word embeddings have been used yet). It uses a simple feed forward neural network to predict labels based on previously given texts. The ability of NLP to learn such simple tasks makes it one of the strongest tools in the field of Artificial/Machine Learning based Chatbots.

demo

Versions:

1. Bag of Words Model:

demo_bag

2. GloVe Model:

demo_glove

Installation:

Use the following command:

git clone https://github.com/AmanPriyanshu/Bavardez.git

To Execute Bavardez:

Use the command:

python bavardez.py

To add additional Intents/Responses to the same:

Use the command:

python intent_initializer.py

The menu for this is self-explanatory allowing most people to use it.

Enter:
        0. If you wish to initialize all the intents.
        1. If you wish to add another intent.
        2. If you wish to edit one of the intents.
Choice:

To Re-Train the Model:

Use the command:

python train.py

Credits:

The bot is loosely based off of the youtube series Chat Bot With PyTorch by Python Engineer