README.md

metadata

title: Semantic Image Search
emoji: 🔍
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 5.16.0
app_file: app.py
pinned: false
license: mit

Semantic Image Search

A semantic image search application that uses knowledge graphs and BLIP-2 for intelligent image retrieval.

Features

• Image captioning using BLIP-2
• Semantic search using ConceptNet knowledge graph
• Adjustable search depth and result count
• Relationship-based image retrieval
• Visual results with explanations

Installation

git clone https://huggingface.co/spaces/[your-username]/semantic-image-search cd semantic-image-search pip install -r requirements.txt

Usage

1. Upload Images:

   • Click "Upload Images" to add images to the system
   • The system will automatically generate captions

2. Search:

   • Enter your search query
   • Adjust the number of results (k) using the slider
   • Set search depth for relationship exploration
   • Click "Search" to retrieve images

Technical Details

• BLIP-2 Model: Salesforce/blip2-opt-2.7b
• Knowledge Graph: ConceptNet
• Interface: Gradio 5.16.0
• Python: 3.9+

Limitations

• Maximum image count: 100
• Supported formats: jpg, jpeg, png
• Free tier resource constraints apply