BinaryVectorDB

# Pre-build Binary Vector Databases Here we ship different pre-build [Binary Vector databases](https://github.com/cohere-ai/BinaryVectorDB). See the Github readme what Binary Vector Database is and why it can save you a lot of memory (and money) when you need to scale to tens or hundred millions of embeddings.  ## Available Datasets We have **Wikipedia** available in all 300+ languages. See [Files](https://huggingface.co/datasets/Cohere/BinaryVectorDB/tree/main) for the different files. These are simple .zip files you can download and unzip locally. Note for the English Wikipedia: English Wikipedia is due to its size distributed as 2 pars: [wikipedia-2023-11-en-part-1.zip](https://huggingface.co/datasets/Cohere/BinaryVectorDB/resolve/main/wikipedia-2023-11-en-part-1.zip) [wikipedia-2023-11-en-part-2.zip](https://huggingface.co/datasets/Cohere/BinaryVectorDB/resolve/main/wikipedia-2023-11-en-part-2.zip) Unzip both to be able to load the Binary Vector Database for the English Wikipedia (42M embeddings).