I was exploring few vector databases and first came across milvus and found working code for it in repo https://github.com/dshvimer/milvus-up-and-running. However, I was only able to make it work for older versions of milvus and connections with new version didnt work which led me towards exploring weaviate vector database. In this repo, I have configured weaviate with similiar use-case, i.e to find similiar images by using vector similiarity search in weaviate. The vectors are computed using Resnet18

Setup Instructions

1. Install docker
2. Run docker-compose build and then docker-compose up -d in the root directory
3. Open http://127.0.0.1:8888/lab on your browser and get the token (or full url with token) from the logs of notebook container by using below commands.
   Use sudo docker ps to list the running containers and get the container id for notebook container and then use sudo docker logs <container_id>, you'll get the url and token at the bottom of logs.
4. To stop the containers use docker-compose down in the same directory.