README.md

metadata

title: E-commerce Product Similarity Engine
emoji: 🔍
colorFrom: blue
colorTo: pink
sdk: streamlit
sdk_version: 1.35.0
app_file: app.py
pinned: false

Product Similarity Engine

This project is a product similarity engine that helps find similar products in an e-commerce catalog using BERT embeddings and cosine similarity.

Features

• Search for products and find similar items instantly.
• Uses precomputed embeddings for fast similarity lookup.
• Clean and simple Streamlit UI for interaction.

Project Structure

. │ app.py │ requirements.txt │ Dockerfile ├─ book_embeddings.pkl └─ books_meta.csv

How It Works

1. Loads product metadata and embeddings.
2. Computes cosine similarity between the selected product and all other products.
3. Returns the top-K most similar items.

Run Locally

pip install -r requirements.txt
streamlit run app.py

Run with Docker
docker build -t product-similarity .
docker run -p 7860:7860 product-similarity

Acknowledgments

SentenceTransformers (for embeddings)

Streamlit for UI

Hugging Face Spaces for hosting