🧠🎨 Controllable Image Generation using with Stable Diffusion and T2I-Adapter

Welcome to the Controllable Image Generation project! This repository explores how to guide image synthesis using T2I-Adapter models with different types of conditioning inputs like edge maps, line poses, and depth maps. Perfect for researchers, creatives, and AI enthusiasts who want precise control over generative models. ✨

---

📁 Notebooks Overview

Notebook	Input Modality	Description
T2IAdapter_EdgeMask.ipynb	🖍️ Edge Mask	Generate images guided by structural outlines (edges). Helps retain object shape and boundary.
T2IAdapter_LinePose.ipynb	🕴️ Line Pose	Condition on pose information to generate people or objects in specific configurations.
T2IAdapter_depth.ipynb	🌄 Depth Map	Leverage spatial depth information to generate 3D-aware scenes.

---

🚀 Getting Started

Follow these steps to run the notebooks:

1. 📥 Clone the repository

git clone https://github.com/meghakalia/Controllable_Image_Generation.git
cd Controllable_Image_Generation

2. 📓 Launch Jupyter Notebook

jupyter notebook

3. ▶️ Run Any Notebook

Choose a notebook based on your use case and run the cells to generate amazing images!

📚 References

T2I-Adapter Paper (arXiv) 📄 Original GitHub Repository (TencentARC) 💻

🤝 Contributing

Feel free to open issues or submit pull requests to improve the project. Feedback is always welcome!