🖼️ Text-to-Image Generator using SDXL (Jupyter Lab)

A professional, research-friendly Text-to-Image generation project built using Stable Diffusion XL (SDXL), PyTorch, and Diffusers, designed to run smoothly inside Jupyter Lab.

This project focuses on:

• Clean architecture
• Efficient GPU/CPU usage
• Simple two-cell workflow
• High-quality image generation

---

🔗 Download SDXL Model (IMPORTANT – Do This First)

👉 Download SDXL Base Model (safetensors) ⬇️ Click Here to Download

⬇️ Download file:

sd_xl_base_1.0.safetensors

📁 After downloading, place the model here:

models/sd_xl_base_1.0.safetensors

⚠️ The notebook will NOT run unless the model is placed correctly inside the models/ folder.

---

📂 Project Structure

text-to-image-sdxl/
│
├── models/
│   └── sd_xl_base_1.0.safetensors
│
├── text_to_image.ipynb
├── README.md
└── requirements.txt (optional)

---

🧠 What is SDXL?

Stable Diffusion XL (SDXL) is a state-of-the-art text-to-image generative model capable of producing:

• Ultra-realistic images
• Cinematic lighting
• High-resolution outputs
• Strong prompt understanding

This project uses local inference, meaning:

• No API cost
• No internet dependency after setup
• Full control over generation

---

🐍 Python Environment Setup (Highly Recommended)

✅ Step 1: Install Python 3.10

Download Python 3.10.x from:

https://www.python.org/downloads/

✔️ Make sure Python is added to PATH during installation.

---

✅ Step 2: Create Virtual Environment (Python 3.10)

python -m venv venv

Activate it:

Windows

venv\Scripts\activate

Linux / macOS

source venv/bin/activate

---

✅ Step 3: Upgrade pip

python -m pip install --upgrade pip

---

🔥 Install Required Libraries

▶️ PyTorch (GPU Recommended)

If you have NVIDIA GPU (CUDA 12.1):

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121

CPU only:

pip install torch torchvision torchaudio

---

▶️ Install Diffusers & Supporting Libraries

pip install diffusers transformers accelerate safetensors

---

▶️ Install Jupyter Lab

pip install jupyterlab ipython

Launch Jupyter:

jupyter lab

---

🧪 Notebook Usage Guide (Very Important)

The notebook is intentionally designed with ONLY TWO MAIN CELLS.

---

🔹 Cell 1: SDXL Model Loader (RUN ONLY ONCE)

📌 Purpose

• Loads the SDXL model into memory
• Detects GPU/CPU automatically
• Applies memory optimizations
• Prepares pipeline for reuse

📌 Rule

🚫 Run this cell ONLY ONE TIME per session

✔️ Re-running it again wastes memory and slows performance.

---

🔹 Cell 2: Image Generation (RUN MULTIPLE TIMES)

📌 Purpose

• Accepts text prompt
• Generates image
• Displays output inside Jupyter

📌 You can safely run this cell multiple times

• Change prompt
• Adjust parameters
• Generate unlimited images

---

📝 Prompt Writing Tips

✔️ Keep prompts descriptive but concise ✔️ Use negative prompts to remove artifacts ✔️ Best resolution: 768 × 768 ✔️ CFG scale between 6 – 8 works best

Example:

Ultra realistic cinematic scene, golden hour lighting,
photorealistic, high detail

---

⚙️ Key Parameters Explained

Parameter	Meaning
width / height	Image resolution
num_inference_steps	More steps = better detail
guidance_scale	Prompt control strength
negative_prompt	Removes unwanted artifacts
torch.no_grad()	Faster & memory-safe inference

---

💻 Hardware Requirements

Minimum

• CPU (works but slower)
• 16 GB RAM

Recommended

• NVIDIA GPU (8 GB+ VRAM)
• CUDA enabled
• SSD storage

---

🚀 Why This Project is Clean & Professional

✅ Local inference (no API dependency) ✅ Jupyter-friendly workflow ✅ Memory-optimized SDXL loading ✅ Clear separation of setup & inference ✅ Easy for demos, research, and presentations

---

📜 Disclaimer

This project is intended for:

• Educational use
• Research
• Demonstrations

Users are responsible for generated content.

---

⭐ Final Note

If you like this project:

• ⭐ Star the repository
• 🍴 Fork it
• 🧠 Experiment with prompts

Happy Generating 🚀