Show HN: Fine-tune an embedding model on your own search logs

Category: ai-ml

Tags: embeddings, fine-tuning, rag, search, openai-compatible, cloud-api

Score: 5.3/10 (Innovation: 5, Technical: 4, Documentation: 7, Utility: 5)

ReOpenly cookbook is a set of runnable Python examples demonstrating how to report search logs, fine-tune an embedding model on your own feedback, and serve it via an OpenAI-compatible API. It targets RAG practitioners who want to improve retrieval with domain-specific embeddings, but relies on ReOpenly's proprietary cloud service, making the repo primarily a client-side cookbook rather than an independent tool.

Target audience: ai engineers, ml engineers, backend developers building RAG applications

Repository: https://reopenly.com/blog/embedding-projects/ · Python · MIT · 1 stars

View on Hacker News