Auto-Labeling Images with Florence 2 and SAM

FIRAS TLILI

ML Engineer

AI Developer

AI Engineer

Hugging Face

Jupyter

Python

Artificial Intelligence

🎯 Auto-Labeling Images with Microsoft Florence 2 and Meta AI SAM

Unlock the potential of Florence 2 & SAM with this step-by-step guide to automatically label images and simplify your AI workflows!

This beginner-friendly project showcases how to leverage Microsoft’s Florence 2 multimodal model and Meta AI's SAM multimodal model for:

🏷️ Automatic data labeling

🧠 Object detection

🖼️ Image classification

🧩 Segmentation

Perfect for AI beginners and enthusiasts alike, these two notebooks provides a clear, well-structured approach to mastering modern AI tools like Hugging Face and Google Colab.

🚀 What You’ll Learn

✅ Set up and use Florence 2 and SAM for various vision tasks

🗂️ Process image data and handle prompts

🔍 Auto-generate annotations for models like YOLOv9

🔧 Use platforms like Roboflow to enhance your pipeline

🧠 Understand the power of multimodal AI

🖼️ Auto Object Detection with Florence-2

🖼️ Auto Image Segmentation with SAM

🛠 Tech Stack

Python

Microsoft Florence 2

Hugging Face

Google Colab

Roboflow

📂 How to Run

# Clone the repo
git clone https://github.com/TLILIFIRAS/Automatic-Data-Annotation-With-LLMs.git 
cd Automatic-Data-Annotation-With-LLMs

# Set up the environment
pip install -r requirements.txt

# Launch the notebook of Data_Labeling with Florence-2
jupyter notebook Automatic_Data_Labeling_with_Florence_2.ipynb

# Launch the notebook of Data_Labeling for Image Segmentation 
jupyter notebook Automatic_Data_Annotation_for_Image_Segmentation.ipynb