Semantic Search

Hey everyone! We just published the Semantic Search article. it’s a gentle intro to one of the key capabilities of language models (representation models in particular). It’s accompanied by code examples in GitHub - cohere-ai/notebooks: Code examples and jupyter notebooks for the Cohere Platform.

This guide uses Cohere’s embeddings and the Annoy library for nearest neighbor retrieval (other options include Faiss, ScaNN, and PyNNDescent).

I like having the major components of the problem space worked out. Since working this out, I’ve been using it to explore online content and communities. Will be posting more about this in the future!

Let me know if you have success or challenges building with this guide.

Happy building!