AI-Powered Floor Planning Solution

✅ Generate floor plans using Fused Stable Diffusion with ControlNet, Gemini with Matplotlib, and Multimodal Stable Diffusion models. ✅ Route and filter user prompts using a NLP filter made using Gemini API. ✅ Built-in model switcher to select desired generator. ✅ Tailored output generation based on prompt engineering and model-specific tuning. ✅ Optimized for fast performance and extensibility.

🏗️ Tech Stack

💻 Web Application

🔹 Frontend

Javascript - Core programming language

React.js + Vite - Frontend framework

Tailwind CSS - Styling library

React Icons - UI enhancement

🔹 Backend

Python - Core programming language

FastAPI / Flask - Web framework

Gemini API - Prompt filtering and refinement

📱 Mobile Application

🔹 Frontend

Typescript - Core programming language

React Native - Frontend framework

Lucide & React Icons - UI enhancement

🔹 Backend

Python - Core programming language

FastAPI / Flask - Web framework

Gemini API - Prompt filtering and refinement

🧠 Models Integrated

🔹 Fused Stable Diffusion and ControlNet Model

Processes user inputs through a specialized NLP filter for image generation.

Utilizes ControlNet for structured guidance of the image creation process.

Implements Stable Diffusion for high-quality base image generation.

Creates initial images based on filtered text prompts.

Applies secondary ControlNet processing to enhance and refine generated images.

Combines text-to-image generation with precise control mechanisms.

Delivers final outputs with improved details and adherence to user specifications.

Enables advanced image manipulation through a dual-phase approach.

🔹 Gemini with Matplotlib Model

Processes user prompts through Google's Gemini model for data visualization tasks.

Uses prompt engineering to generate structured JSON output from natural language requests.

Converts JSON data into Matplotlib visualization code automatically.

Generates customized graphs and charts based on user specifications.

Produces high-quality visualization outputs without requiring manual coding.

Streamlines the data visualization workflow from natural language to final image.

🔹 Multi Stable Diffusion Models

Processes user prompts through a specialized NLP filter for improved understanding

Employs a multi-model architecture to handle varied image generation needs

Converts filtered prompts into a standardized single prompt format

Simultaneously routes prompts to multiple image generation models including:

Stability AI's Stable Diffusion for high-quality general image creation

ZB-Tech Text-to-Image Generator for specialized visual outputs

CloudyAI Text-to-Image Generator for distinct stylistic approaches

Runwayml Stable Diffusion for alternative visual interpretations

Produces multiple output variations from a single user prompt

Enables comparison between different image generation techniques

Provides users with diverse visual interpretations of the same concept

Optimizes prompt handling for compatibility with all connected models

📂 Project Structure

🫚 Directory Tree

HackByte-3.0/
│
├───vite.config.js
├───package.json
├───package-lock.json
├───index.html
├───eslint.config.js
├───README.md
├───LICENSE
├───.gitignore
│
├───src/
│   ├───main.jsx
│   ├───index.css
│   ├───App.jsx
│   ├───App.css
│   └───services/
│   │   └───chatApiService.js
│   │   
│   └───pages/
│   │   ├───StableDiffusionPage.jsx
│   │   ├───ResultsPage.jsx
│   │   ├───MidjourneyPage.jsx
│   │   ├───KnowMorePage.jsx
│   │   ├───HomePage.jsx
│   │   ├───DallEPage.jsx
│   │   ├───ContributorsPage.jsx
│   │   └───ChatPage.jsx
│   │
│   ├───components/
│   │   ├───Welcome.jsx
│   │   ├───PageTransition.jsx
│   │   ├───NavBar.jsx
│   │   ├───MessageList.jsx
│   │   ├───LoadingScreen.jsx
│   │   ├───ChatSidebar.jsx
│   │   └───ChatInput.jsx
│   │
│   └───assets/
│       └───react.svg
│
├───Results/
│   ├───Stable Diffusion/
│   │   ├───Result 5 stable diffusion.png
│   │   ├───Result 4 stable diffusion.png
│   │   ├───Result 3 stable diffusion.png
│   │   ├───Result 2 stable diffusion.png
│   │   └───Result 1 stable diffusion.png
│   │
│   ├───Multi Modal/
│   │   ├───Result 5 multi modal.png
│   │   ├───Result 4 multi modal.png
│   │   ├───Result 3 multi modal.png
│   │   ├───Result 2 multi modal.png
│   │   └───Result 1 multi modal.png
│   │
│   └───Gemini/
│       ├───
│       ├───
│       ├───
│       ├───
│       └───Result 1 gemini.png
│
├───public/
│   ├───LOGO-modified.png
│   ├───profilePhotos/
│   │   ├───bindupautra.jpg
│   │   ├───avinash.jpeg
│   │   └───arunim.jpg
│   │
│   ├───logos/
│   │   ├───stable diffusion and controlnet model logo.jpeg
│   │   ├───multi modal logo.jpeg
│   │   └───gemini matplotlib model logo.jpeg
│   │
│   └───ArchitectureImg/
│       ├───stable diffusion controlnet model architecture.jpg
│       ├───multi modal model architecture.jpg
│       └───Gemini matplotlib model architecture.jpg
│
├───backend/
│   ├───app.py
│   └───requirements.txt
│
├───Notebooks/
│   ├───Readme.md
│   ├───Multi-Modal/
│   │   ├───ZB_Tech_Text_to_Image_Model.ipynb
│   │   ├───ZB_Tech_Text_to_Image_Generator_API.ipynb
│   │   ├───Unoptimized_cloudqi_Text_to_Image_Model.ipynb
│   │   ├───Unified_APIs.ipynb
│   │   ├───URL_Testing.ipynb
│   │   ├───Stability_AI_Stable_Diffusion_Model_API.ipynb
│   │   ├───StabilityAI_Stable_Diffusion_Model.ipynb
│   │   ├───Runwayml_Stable_Diffusion_Model_API.ipynb
│   │   ├───Runwayml_Stable_Diffusion_Model.ipynb
│   │   ├───Optimized_Cloudqi_Text_to_Image_Model.ipynb
│   │   ├───Cloudqi_Text_to_Image_Model_API.ipynb
│   │   └───API_Testing.ipynb
│   │
│   ├───Gemini with Matplotlib Model/
│   │   └───dynamic_floor_planner_ipynb.ipynb
│   │
│   └───Fused Stable Diffusion with ControlNet Model/
│       └───ControlNet_Stable_Diffusion_wandb_(1) (1).ipynb
│
├───Mobile App/
│   ├───tsconfig.json
│   ├───package.json
│   ├───package-lock.json
│   ├───app.json
│   ├───.prettierrc
│   ├───.gitignore
│   ├───hooks/
│   │   └───useFrameworkReady.ts
│   │
│   ├───backend/
│   │   ├───requirements.txt
│   │   └───app.py
│   │
│   ├───assets/
│   │   └───images/
│   │       ├───icon.png
│   │       ├───favicon.png
│   │       ├───Results/
│   │       │   ├───Stable Diffusion/
│   │       │   │   └───Result 1 stable diffusion.png
│   │       │   │
│   │       │   ├───Multi Modal/
│   │       │   │   ├───Result 4 multi modal.png
│   │       │   │   ├───Result 3 multi modal.png
│   │       │   │   ├───Result 2 multi modal.png
│   │       │   │   └───Result 1 multi modal.png
│   │       │   │
│   │       │   └───Gemini/
│   │       │       └───Result 1 gemini.png
│   │       │
│   │       ├───Model Architectures/
│   │       │   ├───Stable Diffusion and ControlNet Architecture.jpg
│   │       │   ├───Multi Modal Architecture .jpg
│   │       │   └───Gemini with matplotlib model architecture.jpg
│   │       │
│   │       └───Contributors/
│   │           ├───Bindupautra Jyotibrat.jpg
│   │           ├───Avinash Kushwaha.jpeg
│   │           ├───Arunim Gogoi.png
│   │           └───Ansh Photo.jpg
│   │
│   └───app/
│       ├───_layout.tsx
│       ├───+not-found.tsx
│       └───(tabs)/
│           ├───results.tsx
│           ├───models.tsx
│           ├───index.tsx
│           ├───contributors.tsx
│           └───_layout.tsx
│
├───Assets/
│   ├───README.md
│   ├───Bindupautra Github Photo.png
│   ├───Arunim Github Photo.png
│   ├───Ansh Github Photo.png
│   └───AK.jpeg
│
└───.github/
     ├───README.md
     └───ISSUE_TEMPLATE/
         ├───feature_request.md
         └───bug_report.md

🔗 Links to Directories

Gemini with Matplotlib Model/

Fused Stable Diffusion with ControlNet Model/

🚀 Deployment Status

🔹 Frontend Deployment

👉 Live Frontend Preview

⚠️ Note: Backend deployment is currently under development. Please follow the setup instructions below to run the full application locally.

🚀 Installation & Setup

Clone the repository

git clone https://github.com/Jyotibrat/HackByte-3.0.git

💻 For Web Application

🖥️ Frontend Setup

Install the required Node Modules.

npm install

If Node packages require funding.

npm fund

If there are any Vulnerabilities.

npm audit fix

If the Vulnerabilities still not resolved.

npm audit fix --force

Create a .env file and the following lines:

VITE_BACKEND_URL = http://localhost:8000/api/generate

🧠 Backend Setup

Move to the backend Directory.

cd backend

Create a Virtual Environment.

python -m venv env

Activate the Virtual Environment.

env\Scripts\activate

Download the required packages for the Backend.

pip install -r requirements.txt

Generate a GEMINI 1.5-flash API from Google AI Studio.

On the NavBar on the top click on Get API key button.

In the next page, click on Create API key button on the right side.

After the key has been generated copy it.

Now, Create a .env file in the backend Directory add the following lines:

GEMINI_API_KEY=REPLACE_WITH_YOUR_API_KEY
PORT=5000
DEBUG=True

Move to the Notebooks Directory.

cd Notebooks

Move to the Gemini with Matplotlib Model Directory.

cd "Gemini with Matplotlib Model"

Go to Google Colab and Click on Upload button.

Upload the Notebook dynamic_floor_planner_ipynb.ipynb on Google Colab.

Add the GEMINI 1.5-flash API that was generated earlier here:

genai.configure(api_key="PASTE GEMINI API HERE")

Go to ngrok and sign up.

On the left side, click on the Your Authtoken button.

Copy the Authtoken and paste it here in the Notebook:

ngrok.set_auth_token("YOUR_GENERATED_AUTHTOKEN")

Run the cells of the Notebook in top to bottom manner.

After the running of the cells in the Notebook is completed, a ngrok url will be generated.

The ngrok url would look like this:

"https://4539-35-245-46-130.ngrok-free.app"

Copy the url and paste it in app.py in the backend Directory like this:

"floor_gpt": {
        "name": "floor_gpt",
        "urls": ["PASTE_THE_URL_HERE"]
    }

For running the second model, move to Fused Stable Diffusion with ControlNet Model in the Notebooks Directory.

cd Notebooks
cd "Fused Stable Diffusion with ControlNet Model"

Upload the Notebook ControlNet_Stable_Diffusion_wandb API.ipynb on Google Colab.

On the Navbar, click the Runtime button.

From the drop down menu, click the Change runtime type button.

From the options, select the T4 GPU and click on Save button.

Run the cells of the Notebook in top to bottom manner.

After the running of the cells in the Notebook is completed, a ngrok url will be generated.

The ngrok url would look like this:

"https://4539-35-245-46-130.ngrok-free.app"

Copy the url and paste it in app.py in the backend Directory like this:

"stable_diffusion": {
        "name": "stable_diffusion",
        "urls": ["https://2a16-34-125-2-19.ngrok-free.app/"]
    },

For running the second model, move to Multi-Modal in the Notebooks Directory.

cd Notebooks
cd "Multi-Modal"

Upload the Notebooks Cloudqi_Text_to_Image_Model_API.ipynb, Runwayml_Stable_Diffusion_Model_API.ipynb, Stability_AI_Stable_Diffusion_Model_API.ipynb and ZB_Tech_Text_to_Image_Generator_API.ipynb on Google Colab

⚠️ Important: Upload each notebook using a different Google account to get free GPU access per notebook (one GPU per account on the free tier). If you have Google Colab Pro or Premium, you can ignore this step.

In each Notebook:

On the Navbar, click the Runtime button.

From the drop-down menu, click the Change runtime type button.

From the options, select the T4 GPU and click on Save button.

Run the cells of the Notebook in top to bottom manner.

After the running of the cells in the Notebook is completed, a ngrok url will be generated.

The ngrok url would look like this:

After the url has been generated, run the second last cell.

Copy the urls and paste it in app.py in the backend Directory like this:

"multi_modal": {
        "name": "multi_modal",
        "urls": [
            "https://2a16-34-125-2-19.ngrok-free.app/",
            "https://2127-34-16-255-178.ngrok-free.app/",
            "https://third-ngrok-url.ngrok-free.app/",
            "https://fourth-ngrok-url.ngrok-free.app/"
        ]
    },

🚀 Run Application

Go to the root Directory and run the command:

npm run dev

Open your browser and navigate to http://localhost:5173 to see the application.

Go to backend Directory and make sure the Virtual Environment is activate.

cd backend
env\Scripts\activate

To run the Backend Server, run the command:

python app.py

✅ The Web Application is running successfully.

📱 For Mobile Application

Go to the Mobile App Directory.

cd "Mobile App"

🖥️ Frontend Setup

Install the required Node Modules.

npm install

If Node packages require funding.

npm fund

If there are any Vulnerabilities.

npm audit fix

If the Vulnerabilities still not resolved.

npm audit fix --force

Go to the app Directory and then to (tabs) Directory.

cd "app/(tabs)"

Open the file index.tsx.

In the file, add the ngrok URL that exposes the http://127.0.0.1:5000/ port of the Flask backend.

const API_URL = 'PASTE_THE_URL_HERE';

🧠 Backend Setup

Move to the backend Directory.

cd backend

Create a Virtual Environment.

python -m venv env

Activate the Virtual Environment.

env\Scripts\activate

Download the required packages for the Backend.

pip install -r requirements.txt

Generate a GEMINI 1.5-flash API from Google AI Studio

On the NavBar on the top click on Get API key button.

In the next page, click on Create API key button on the right side.

` After the key has been generated copy it.

Now, Create a .env file in the backend Directory add the following lines:

GEMINI_API_KEY=REPLACE_WITH_YOUR_API_KEY
PORT=5000
DEBUG=True

Move to the Notebooks Directory.

cd Notebooks

Move to the Gemini with Matplotlib Model Directory.

cd "Gemini with Matplotlib Model"

Go to Google Colab and Click on Upload button.

Upload the Notebook dynamic_floor_planner_ipynb.ipynb on Google Colab.

Add the GEMINI 1.5-flash API that was generated earlier here:

genai.configure(api_key="PASTE GEMINI API HERE")

Go to ngrok and sign up.

On the left side, click on the Your Authtoken button.

Copy the Authtoken and paste it here in the Notebook:

ngrok.set_auth_token("YOUR_GENERATED_AUTHTOKEN")

Run the cells of the Notebook in top to bottom manner.

After the running of the cells in the Notebook is completed, a ngrok url will be generated.

The ngrok url would look like this:

"https://4539-35-245-46-130.ngrok-free.app"

Copy the url and paste it in app.py in the backend Directory like this:

"floor_gpt": {
        "name": "floor_gpt",
        "urls": ["PASTE_THE_URL_HERE"]
    }

For running the second model, move to Fused Stable Diffusion with ControlNet Model in the Notebooks Directory.

cd Notebooks
cd "Fused Stable Diffusion with ControlNet Model"

Upload the Notebook ControlNet_Stable_Diffusion_wandb API.ipynb on Google Colab

On the Navbar, click the Runtime button.

From the drop down menu, click the Change runtime type button.

From the options, select the T4 GPU and click on Save button.

Run the cells of the Notebook in top to bottom manner.

After the running of the cells in the Notebook is completed, a ngrok url will be generated.

The ngrok url would look like this:

"https://4539-35-245-46-130.ngrok-free.app"

Copy the url and paste it in app.py in the backend Directory like this:

"stable_diffusion": {
        "name": "stable_diffusion",
        "urls": ["https://2a16-34-125-2-19.ngrok-free.app/"]
    },

For running the second model, move to Multi-Modal in the Notebooks Directory.

cd Notebooks
cd "Multi-Modal"

In each Notebook:

On the Navbar, click the Runtime button.

From the drop-down menu, click the Change runtime type button.

From the options, select the T4 GPU and click on Save button.

Run the cells of the Notebook in top to bottom manner.

After the running of the cells in the Notebook is completed, a ngrok url will be generated.

The ngrok url would look like this:

After the url has been generated, run the second last cell.

Copy the urls and paste it in app.py in the backend Directory like this:

"multi_modal": {
        "name": "multi_modal",
        "urls": [
            "https://2a16-34-125-2-19.ngrok-free.app/",
            "https://2127-34-16-255-178.ngrok-free.app/",
            "https://third-ngrok-url.ngrok-free.app/",
            "https://fourth-ngrok-url.ngrok-free.app/"
        ]
    },

Download the ngrok application on your local device:

Windows → Download for Windows

Linux → Download for Linux

macOS → Download for macOS

Extract the downloaded zip file to a directory of your choice.

Start the ngrok application and provide the necessary permissions.

🚀 Run Application

Go to the Mobile App Directory and run the command:

npx expo start

Open your browser and navigate to http://localhost:8081 to see the application.

Download the Expo Go Mobile App.

In the app, there will be 2 options:

Enter URL manually where you can enter the url generated in the terminal when the frontend server starts.

Scan QR code where you can scan the QR code generated in the terminal when the frontend server starts.

Now, the app starts building in the mobile.

Go to backend Directory and make sure the Virtual Environment is activate.

cd backend
env\Scripts\activate

To run the Backend Server, run the command:

python app.py

When the ngrok application starts it starts in the command line where we need to run the following command to expose the running Flask Backend server:

ngrok http 5000

This will generate a ngrok url like this:

ngrok                                                                                                   (Ctrl+C to quit)                                                                                                                        �  Route traffic by anything: https://ngrok.com/r/iep                                                                                                                                                                                           Session Status                online                                                                                    Account                       YOUR_NAME (Plan: Free)                                                        Update                        update available (version 3.22.1, Ctrl-U to update)                                       Version                       3.22.0                                                                                    Region                        India (in)                                                                                Latency                       710ms                                                                                     Web Interface                 http://127.0.0.1:4040                                                                     Forwarding                    https://218a-139-167-143-182.ngrok-free.app -> http://localhost:5000                                                                                                                                              Connections                   ttl     opn     rt1     rt5     p50     p90                                                                             0       0       0.00    0.00    0.00    0.00

Copy the url and paste it here in the index.tsx:

const API_URL = 'PASTE_THE_URL_HERE';

✅ The Mobile Application is running successfully.

👥 Contributors

This project was made possible by the contributions of these amazing individuals:

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.

🏆 Acknowledgment

This project is part of HackByte 3.0 conducted by IIITDM Jabalpur.

Like this project

Completed work

Posted May 25, 2025

AI solution for generating architectural floor plans using generative models.

Likes

Views