Fine-tuning and implementing retrievers in open-domain Q&A

Hi all, I've just published another article focusing on fine-tuning a retriever model for open-domain question-answering. The retriever is a big component of the open-domain QA pipeline, allowing us to retrieve relevant contexts from a vector database, which then help us answer a query (article includes vector db <> retriever setup too)

Let me know if you have any questions or feedback, thanks!

fine-tuning languagetechnology