ba59f077a2* inputs * agents * chains * custom-component * align-admonitions-in-custom * data-and-embeddings * experimental * helpers * memories * model_specs * outputs * prompts * retrievers * textsplitter * tools * utilities * vector-stores
24 lines
1.3 KiB
Text
24 lines
1.3 KiB
Text
import Admonition from '@theme/Admonition';
|
|
|
|
# Retrievers
|
|
|
|
<Admonition type="caution" icon="🚧" title="ZONE UNDER CONSTRUCTION">
|
|
<p>
|
|
We appreciate your patience as we enhance our documentation. It may have some imperfections. Please share your feedback or report issues to help us improve. 🛠️📝
|
|
</p>
|
|
</Admonition>
|
|
|
|
A retriever is an interface that returns documents in response to an unstructured query. It's broader than a vector store because it doesn't need to store documents; it only needs to retrieve them.
|
|
|
|
---
|
|
|
|
### MultiQueryRetriever
|
|
|
|
The `MultiQueryRetriever` automates generating multiple queries, retrieves relevant documents for each query, and aggregates the results. This method improves retrieval effectiveness and addresses the limitations of traditional distance-based methods.
|
|
|
|
**Parameters**
|
|
|
|
- **LLM:** Specifies the language model used in the `MultiQueryRetriever`.
|
|
- **Prompt:** Defines a schema for the LLM.
|
|
- **Retriever:** Identifies the retriever that fetches documents.
|
|
- **parser_key:** Specifies the key or attribute name of the parsed output for retrieval. By default, it's set to `lines`, meaning the output from the language model is split into separate lines of text. This allows the retriever to fetch documents relevant to each line of text.
|