🚀 AutoPrompt2Image

LLM-powered prompt optimization for Stable Diffusion with LoRA-enhanced image generation.

Turn natural language into high-quality AI-generated images.

AutoPrompt2Image is an end-to-end pipeline that leverages a LLaMA-based language model to transform raw user input into high-quality, diffusion-ready prompts, and generates images using a LoRA fine-tuned Stable Diffusion model.

---

✨ Features

• 🔹 Automatic Prompt Engineering via LLaMA
• 🔹 LoRA Fine-tuned Stable Diffusion for style consistency
• 🔹 End-to-End Text-to-Image Pipeline
• 🔹 Custom Dataset Training Support
• 🔹 Modular Design (easy to extend or replace components)

---

🧩 Pipeline

User Input
   ↓
LLaMA (Prompt Optimization)
   ↓
Optimized Prompt
   ↓
LoRA Stable Diffusion
   ↓
Generated Image

⚙️ Installation

git clone https://github.com/J-damn649/AutoPrompt2Image.git
cd AutoPrompt2Image

pip install -r requirements_genimage.txt
pip install -r requirements_gendata.txt

---

🚀 Usage

python scripts/main.py --prompt "a cat in van gogh style"

---

📁 Project Structure

AutoPrompt2Image/
│── scripts/        # inference / entry scripts
│── models/         # model checkpoints (not included)
│── trainer/        # training code (LoRA, SFT)
│── dataset/        # training data (not included)
│── outs/           # generated images             
│── requirements_gendata.txt
│── requirements_genimage.txt
│── README.md

---

🧠 How It Works

1. User provides a natural language description
2. LLaMA refines it into a structured prompt
3. Prompt is fed into Stable Diffusion
4. LoRA enhances style and consistency
5. Final image is generated

---

🙌 Acknowledgements

• Stable Diffusion
• HuggingFace Diffusers
• LLaMA / Transformers
• PEFT (LoRA)

---

⭐ If you like this project

Give it a star ⭐ on GitHub!