/lens

Primary LanguagePython

Setup

1. Install the Weaviate python client

pip install weaviate-client

2. Run weaviate with clip module in docker

docker-compose up -d

3. Create the schema

python create_schema.py

4. Download your dataset images into dataset_images folder

This is a manual step

5. Insert the images into Weaviate

python import_images.sh

Run

Query by text

Edit query_by_text.py with your desired text and then run: python query_by_text.py

Query by image

Add an image to query_images folder, and point to the image in query_by_image.py, then run: python query_by_image.py

You can also drop the schema to start again

./drop_schema.sh