Memory & evals · Project

Ragas

Ragas scores retrieval and generation quality and builds synthetic evaluation datasets for LLM applications.

+ Follow

Builder facts

  • License. The Python package uses the Apache 2.0 license. Source ↗
  • Model access. Model-backed metrics use the configured evaluator, such as an OpenAI client with its own credentials. Source ↗
  • Retrieval ranking. Context precision measures how highly a retriever ranks relevant chunks. Available variants use a reference answer, a generated response, or reference contexts without a model judge. Source ↗
  • Answer grounding. Faithfulness measures the fraction of response claims supported by the retrieved context. Source ↗
  • Synthetic datasets. The test-set generator builds a knowledge graph from documents and uses a generator model and embeddings to create single-hop and multi-hop examples. Source ↗

Sources checked

Publisher update ·

Ragas 0.4.4 ↗

Made Vertex AI imports optional to restore fresh installations with newer langchain-community versions and corrected request compatibility for newer Anthropic and OpenAI model IDs.

Change timeline

0 stories

No verified releases in this timeline yet.