CHATBOT

Here we use various NLP concepts to make a chatbot . This is customisable by changing prompts in the 'intents.json' file.

The steps and concepts used are:

1. NLP preprocessing pipeline

• Tokenise the sentences/ prompts given to tokens/words.
• Lowercase all the words and them stem them with stemming.
• Exclusion of any punctuation characters.
• Use bag of words to change the words to '0's and '1's.

2. Create a training set

3. Pytorch model building and training

4. Save/load model and implement the chat

5. Create a GUI

Preview: