pal2004/Vlm-search

🔍 VLM Cloud Image Search

Search through images from cloud sources using AI-powered content recognition

This application uses OpenAI's CLIP (Contrastive Language-Image Pre-training) model to understand image content and find matches from various cloud sources.

✨ Features

• —🌐 Cloud Image Fetching: Load images from URLs, Google Drive, local files
• —🤖 AI Content Search: Describe what you're looking for in natural language
• —📁 Filename Search: Search by exact or partial filename
• —📊 Similarity Scoring: See how confident the AI is about matches
• —🖼️ Image Preview: View results immediately
• —⚡ Smart Caching: Images cached for faster repeated searches
• —📱 Responsive Design: Works on desktop and mobile

🚀 How to Use

1. Add Image Sources

In the "Image Sources" box, add one source per line:

https://example.com/banana.jpg
https://drive.google.com/file/d/1ABC123DEF456/view
/path/to/local/cat.jpg
https://picsum.photos/400/300

2. Search Images

Content Search (AI)

• —Type descriptions like: "yellow fruit", "cute animal", "red car"
• —The AI will analyze image content and find matches
• —See similarity scores and confidence levels

Filename Search

• —Type exact filenames: "banana.jpg", "cat.png"
• —Or partial names: "banana", "cat"
• —Perfect for finding specific files

3. View Results

• —See the matched image
• —Check similarity score and confidence
• —View source information

🌐 Supported Cloud Sources

• —Direct URLs: https://example.com/image.jpg
• —Google Drive: https://drive.google.com/file/d/FILE_ID/view
• —Local Files: /path/to/image.jpg
• —Base64: data:image/jpeg;base64,...

🛠️ Technical Details

• —Model: OpenAI CLIP ViT-Base-Patch16
• —Framework: Gradio
• —Backend: PyTorch + Transformers
• —Deployment: Hugging Face Spaces
• —Caching: Local image caching for performance

🔧 Local Development

pip install -r requirements.txt
python app.py

📄 License

MIT License - feel free to use and modify!