Gen AI Chat Bot Python Project using Sentence transformer model all-MiniLM-L6-v2

Step 1 –

Import libraries needed for text encoding, similarity calculation, and chatbot interface.

SentenceTransformer –

Converts sentences into numerical vectors so machines can

understand the meaning of text.

scikit-learn (cosine_similarity) –

Calculates similarity between two text vectors to find how closely related they are.

NumPy –

Used for numerical computing, arrays, matrices, and fast mathematical operations.

Pandas –

Used for handling datasets in table format such as reading data, filtering,

cleaning, and manipulating rows and columns.

ipywidgets –

Creates interactive UI elements like text boxes, buttons,

sliders, and dropdowns in notebook environments.

IPython.display –

Displays widgets, HTML, images, and formatted output inside notebooks.

Step 2 –

Loading the model – SentenceTransformer loads the model "all-MiniLM-L6-v2" which converts sentences into numerical embeddings.

Step 3 – Creating product data –

A dictionary is created with product names and product descriptions

where each product has a corresponding description explaining its features.

Step 4 –

Converting dictionary to table –

Pandas converts the dictionary into a DataFrame table with columns and rows

where each row represents a product and its description.

Step 5 –

Combining product and description – The product column and description column

are merged into one combined text column so the model understands the

product name and description together.

Step 6 –

Converting text to vectors – The combined text is passed into the SentenceTransformer

model which converts each sentence into numerical embeddings

representing semantic meaning.

Step 7 –

Creating conversational message lists – Lists are created for greetings,

courtesy messages, goodbye messages, and “how are you” type messages

so the chatbot can respond naturally.

Step 8 –

Welcome message – A welcome message is created using HTML

formatting that introduces the chatbot as an AI product specialist

when the chat interface loads.

Step 9 –

Creating chat box – An HTML chat area is created using ipywidgets which

acts as the conversation window between the user and the chatbot.

Step 10 –

Creating input box – A text input box is created where the user types product

questions and it displays a placeholder prompt.

Step 11 –

Creating send button – A button is created so the user can send their message to

the chatbot.

Step 12 –

Chatbot response logic – A function processes the user message, converts it

to embeddings, compares similarity with product vectors using cosine similarity,

finds the best match, and returns the product recommendation or conversational response.

Step 13 –

Chat message handling function – Another function reads the user message

from the input box, sends it to the chatbot response function, stores

both the user message and bot reply in chat history, updates the chat interface,

and clears the input box.

Step 14 –

send_button.on_click(send_message) – Clicking the send button triggers

the function that processes and sends the message.

Step 15 –

input_box.on_submit(send_message) – Pressing Enter in

the input box also triggers the same message sending function.

Step 16 –

Display interface –

The chat box, input box, and send button are displayed together to

create the interactive chatbot UI.

Final Output –

A simple Small Language Model AI chatbot that understands

product questions, compares semantic meaning using embeddings, recommends

relevant products, and responds conversationally to greetings and common messages.