Rankers
Rankers are a group of components that order documents by given criteria. Their goal is to improve your document retrieval results.
| AmazonBedrockRanker | Ranks documents based on their similarity to the query using Amazon Bedrock models. |
| CohereRanker | Ranks documents based on their similarity to the query using Cohere rerank models. |
| FastembedRanker | Ranks documents based on their similarity to the query using cross-encoder models supported by FastEmbed. |
| HuggingFaceTEIRanker | Ranks documents based on their similarity to the query using a Text Embeddings Inference (TEI) API endpoint. |
| JinaRanker | Ranks documents based on their similarity to the query using Jina AI models. |
| LostInTheMiddleRanker | Positions the most relevant documents at the beginning and at the end of the resulting list while placing the least relevant documents in the middle, based on a research paper. |
| MetaFieldRanker | A lightweight Ranker that orders documents based on a specific metadata field value. |
| MetaFieldGroupingRanker | Reorders the documents by grouping them based on metadata keys. |
| NvidiaRanker | Ranks documents using large-language models from NVIDIA NIMs . |
| TransformersSimilarityRanker | A legacy version of SentenceTransformersSimilarityRanker. |