Loading...
Skip to main content

Interface: Embedding

batteries/docs.Embedding

A function that maps strings to vectors encoding their semantic meaning. Often this is based on the same function used to transform text for an LLM's transformers, though this isn't required. Alse see https://docs.ai-jsx.com/guides/brand-new#semantic-similarity-embeddings.

Implemented by​

Methods​

embed​

▸ embed(text): Promise<number[]>

Parameters​

NameType
textstring

Returns​

Promise<number[]>

Defined in​

ai-jsx/src/batteries/docs.tsx:318

embedBatch​

▸ embedBatch(chunks): Promise<number[][]>

Parameters​

NameType
chunksstring[]

Returns​

Promise<number[][]>

Defined in​

ai-jsx/src/batteries/docs.tsx:319