Class: CustomHuggingFaceTransformersEmbeddings
helpers/langchain/custom-hugging-face-transformers-embeddings.CustomHuggingFaceTransformersEmbeddings
Custom wrapper around HuggingFaceTransformersEmbeddings to support text truncation to a specified maximum number of tokens before embedding. This can be useful when working with models that have a fixed maximum input size, but produce better results when you use a lower input size as the maximum.
Hierarchy
-
HuggingFaceTransformersEmbeddings↳
CustomHuggingFaceTransformersEmbeddings
Table of contents
Constructors
Methods
Constructors
constructor
• new CustomHuggingFaceTransformersEmbeddings(fields?): CustomHuggingFaceTransformersEmbeddings
Parameters
| Name | Type |
|---|---|
fields? | Partial<HuggingFaceTransformersEmbeddingsParams> & { maxTokens?: number } |
Returns
CustomHuggingFaceTransformersEmbeddings
Overrides
HuggingFaceTransformersEmbeddings.constructor
Defined in
Methods
embedDocuments
▸ embedDocuments(texts): Promise<number[][]>
Embeds multiple documents, optionally truncating each to a maximum token length.
Parameters
| Name | Type | Description |
|---|---|---|
texts | string[] | The array of text strings to embed. |
Returns
Promise<number[][]>
A Promise that resolves to a two-dimensional array of embeddings, with each sub-array representing the embedding of one input text.
Overrides
HuggingFaceTransformersEmbeddings.embedDocuments
Defined in
embedQuery
▸ embedQuery(text): Promise<number[]>
Embeds a single query, optionally truncating it to a maximum token length.
Parameters
| Name | Type | Description |
|---|---|---|
text | string | The text string to embed. |
Returns
Promise<number[]>
A Promise that resolves to an array representing the embedding of the input text.
Overrides
HuggingFaceTransformersEmbeddings.embedQuery