CLIP-powered Semantic Search Engine using Pinecone, FiftyOne, and Streamlit

This repository will allow you to generate text embeddings for the COCO2017 dataset, as well as deploy a Streamlit webapp search engine on your local machine.

If you want to check out the deployed app, click here. The app works best on Safari, Mozilla, or DuckDuckGo. You might experience bugs on Chromium-based browsers.

Uploading the Data on your Pinecone

1. Clone this repository

2. Run pip install -r requirements.txt

3. Make sure you have a Pinecone account and API Key. You can find this on your console.

4. Create a file in your directory called config.py, in which you create a variable called PINECONE_KEY = your_key_here

5. Run the data_upload.ipynb notebook

Great! You should have an Index now on your Pinecone console, called clip-image-search.

Running the webapp locally

1. Install Streamlit on your machine if you don't have it

2. In the .streamlit folder, create a secrets.toml file

3. Insert your Pinecone key, such as PINECONE_KEY = your_key_here

4. Run streamlit run streamlit_app.py

The app should be on your localhost! Have fun searching :)