Create a new Embedding Model Configuration

POST 
/api/v1/embedding-model-configs

Create a new embedding model configuration within a specified project.

Request

Query Parameters

project_id uuidrequired

Cookie Parameters

session any

application/json

Body

required

name Name (string)required

The name of the embedding model config.

embedding_config

object

required

The embedding configuration for the embedding model config.

oneOf

AzureOpenAIEmbeddingConfig

type Type (string)

Type of the embedding model.

Possible values: [AZURE_EMBEDDING]

Default value: AZURE_EMBEDDING

component

object

Configuration for the Azure OpenAI embedding model.

model_name Model Name (string)

The name of the OpenAI embedding model.

Default value: text-embedding-ada-002

embed_batch_size Embed Batch Size (integer)

The batch size for embedding calls.

Possible values: > 0 and <= 2048

Default value: 10

num_workers

object

The number of workers to use for async embedding calls.

anyOf

MOD1

integer

additional_kwargs object

Additional kwargs for the OpenAI API.

api_key

object

The OpenAI API key.

anyOf

MOD1

string

api_base Api Base (string)

The base URL for Azure deployment.

Default value:

api_version Api Version (string)

The version for Azure OpenAI API.

Default value:

max_retries Max Retries (integer)

Maximum number of retries.

Default value: 10

timeout Timeout (number)

Timeout for each request.

Default value: 60

default_headers

object

The default headers for API requests.

anyOf

MOD1

property name* string

reuse_client Reuse Client (boolean)

Reuse the OpenAI client between requests. When doing anything with large volumes of async API calls, setting this to false can improve stability.

Default value: true

dimensions

object

The number of dimensions on the output embedding vectors. Works only with v3 embedding models.

anyOf

MOD1

integer

azure_endpoint

object

The Azure endpoint to use.

anyOf

MOD1

string

azure_deployment

object

The Azure deployment to use.

anyOf

MOD1

string

class_name Class Name (string)

Default value: AzureOpenAIEmbedding

CohereEmbeddingConfig

type Type (string)

Type of the embedding model.

Possible values: [COHERE_EMBEDDING]

Default value: COHERE_EMBEDDING

component

object

Configuration for the Cohere embedding model.

model_name Model Name (string)

The modelId of the Cohere model to use.

Default value: embed-english-v3.0

embed_batch_size Embed Batch Size (integer)

The batch size for embedding calls.

Possible values: > 0 and <= 2048

Default value: 10

num_workers

object

The number of workers to use for async embedding calls.

anyOf

MOD1

integer

api_key

object

required

The Cohere API key.

anyOf

MOD1

string

truncate Truncate (string)

Truncation type - START/ END/ NONE

Default value: END

input_type

object

Model Input type. If not provided, search_document and search_query are used when needed.

anyOf

MOD1

string

embedding_type Embedding Type (string)

Embedding type. If not provided float embedding_type is used when needed.

Default value: float

class_name Class Name (string)

Default value: CohereEmbedding

GeminiEmbeddingConfig

type Type (string)

Type of the embedding model.

Possible values: [GEMINI_EMBEDDING]

Default value: GEMINI_EMBEDDING

component

object

Configuration for the Gemini embedding model.

model_name Model Name (string)

The modelId of the Gemini model to use.

Default value: models/embedding-001

embed_batch_size Embed Batch Size (integer)

The batch size for embedding calls.

Possible values: > 0 and <= 2048

Default value: 10

num_workers

object

The number of workers to use for async embedding calls.

anyOf

MOD1

integer

title

object

Title is only applicable for retrieval_document tasks, and is used to represent a document title. For other tasks, title is invalid.

anyOf

MOD1

string

task_type

object

The task for embedding model.

anyOf

MOD1

string

api_key

object

API key to access the model. Defaults to None.

anyOf

MOD1

string

api_base

object

API base to access the model. Defaults to None.

anyOf

MOD1

string

transport

object

Transport to access the model. Defaults to None.

anyOf

MOD1

string

class_name Class Name (string)

Default value: GeminiEmbedding

HuggingFaceInferenceAPIEmbeddingConfig

type Type (string)

Type of the embedding model.

Possible values: [HUGGINGFACE_API_EMBEDDING]

Default value: HUGGINGFACE_API_EMBEDDING

component

object

Configuration for the HuggingFace Inference API embedding model.

model_name

object

Hugging Face model name. If None, the task will be used.

anyOf

MOD1

string

embed_batch_size Embed Batch Size (integer)

The batch size for embedding calls.

Possible values: > 0 and <= 2048

Default value: 10

num_workers

object

The number of workers to use for async embedding calls.

anyOf

MOD1

integer

pooling

object

Pooling strategy. If None, the model's default pooling is used.

anyOf

Pooling

Enum of possible pooling choices with pooling behaviors.

string

Possible values: [cls, mean, last]

query_instruction

object

Instruction to prepend during query embedding.

anyOf

MOD1

string

text_instruction

object

Instruction to prepend during text embedding.

anyOf

MOD1

string

token

object

Hugging Face token. Will default to the locally saved token. Pass token=False if you don’t want to send your token to the server.

anyOf

MOD1

string

MOD2

boolean

timeout

object

The maximum number of seconds to wait for a response from the server. Loading a new model in Inference API can take up to several minutes. Defaults to None, meaning it will loop until the server is available.

anyOf

MOD1

number

headers

object

Additional headers to send to the server. By default only the authorization and user-agent headers are sent. Values in this dictionary will override the default values.

anyOf

MOD1

property name* string

cookies

object

Additional cookies to send to the server.

anyOf

MOD1

property name* string

task

object

Optional task to pick Hugging Face's recommended model, used when model_name is left as default of None.

anyOf

MOD1

string

class_name Class Name (string)

Default value: HuggingFaceInferenceAPIEmbedding

OpenAIEmbeddingConfig

type Type (string)

Type of the embedding model.

Possible values: [OPENAI_EMBEDDING]

Default value: OPENAI_EMBEDDING

component

object

Configuration for the OpenAI embedding model.

model_name Model Name (string)

The name of the OpenAI embedding model.

Default value: text-embedding-ada-002

embed_batch_size Embed Batch Size (integer)

The batch size for embedding calls.

Possible values: > 0 and <= 2048

Default value: 10

num_workers

object

The number of workers to use for async embedding calls.

anyOf

MOD1

integer

additional_kwargs object

Additional kwargs for the OpenAI API.

api_key

object

The OpenAI API key.

anyOf

MOD1

string

api_base

object

The base URL for OpenAI API.

anyOf

MOD1

string

api_version

object

The version for OpenAI API.

anyOf

MOD1

string

max_retries Max Retries (integer)

Maximum number of retries.

Default value: 10

timeout Timeout (number)

Timeout for each request.

Default value: 60

default_headers

object

The default headers for API requests.

anyOf

MOD1

property name* string

reuse_client Reuse Client (boolean)

Reuse the OpenAI client between requests. When doing anything with large volumes of async API calls, setting this to false can improve stability.

Default value: true

dimensions

object

The number of dimensions on the output embedding vectors. Works only with v3 embedding models.

anyOf

MOD1

integer

class_name Class Name (string)

Default value: OpenAIEmbedding

VertexAIEmbeddingConfig

type Type (string)

Type of the embedding model.

Possible values: [VERTEXAI_EMBEDDING]

Default value: VERTEXAI_EMBEDDING

component

object

Configuration for the VertexAI embedding model.

model_name Model Name (string)

The modelId of the VertexAI model to use.

Default value: textembedding-gecko@003

embed_batch_size Embed Batch Size (integer)

The batch size for embedding calls.

Possible values: > 0 and <= 2048

Default value: 10

num_workers

object

The number of workers to use for async embedding calls.

anyOf

MOD1

integer

location Location (string)required

The default location to use when making API calls.

project Project (string)required

The default GCP project to use when making Vertex API calls.

embed_mode VertexEmbeddingMode (string)

The embedding mode to use.

Possible values: [default, classification, clustering, similarity, retrieval]

Default value: retrieval

additional_kwargs object

Additional kwargs for the Vertex.

client_email

object

required

The client email for the VertexAI credentials.

anyOf

MOD1

string

token_uri

object

required

The token URI for the VertexAI credentials.

anyOf

MOD1

string

private_key_id

object

required

The private key ID for the VertexAI credentials.

anyOf

MOD1

string

private_key

object

required

The private key for the VertexAI credentials.

anyOf

MOD1

string

class_name Class Name (string)

Default value: VertexTextEmbedding

BedrockEmbeddingConfig

type Type (string)

Type of the embedding model.

Possible values: [BEDROCK_EMBEDDING]

Default value: BEDROCK_EMBEDDING

component

object

Configuration for the Bedrock embedding model.

model_name Model Name (string)

The modelId of the Bedrock model to use.

Default value: amazon.titan-embed-text-v1

embed_batch_size Embed Batch Size (integer)

The batch size for embedding calls.

Possible values: > 0 and <= 2048

Default value: 10

num_workers

object

The number of workers to use for async embedding calls.

anyOf

MOD1

integer

profile_name

object

The name of aws profile to use. If not given, then the default profile is used.

anyOf

MOD1

string

aws_access_key_id

object

AWS Access Key ID to use

anyOf

MOD1

string

aws_secret_access_key

object

AWS Secret Access Key to use

anyOf

MOD1

string

aws_session_token

object

AWS Session Token to use

anyOf

MOD1

string

region_name

object

AWS region name to use. Uses region configured in AWS CLI if not passed

anyOf

MOD1

string

max_retries Max Retries (integer)

The maximum number of API retries.

Possible values: > 0

Default value: 10

timeout Timeout (number)

The timeout for the Bedrock API request in seconds. It will be used for both connect and read timeouts.

Default value: 60

additional_kwargs object

Additional kwargs for the bedrock client.

class_name Class Name (string)

Default value: BedrockEmbedding

Responses

201

Successful Response

application/json

Schema

Schema

any

422

Validation Error

application/json

Schema

Schema

detail

object[]

Array [

loc

object[]

required

Array [

anyOf

MOD1

string

MOD2

integer

]

msg Message (string)required

type Error Type (string)required

]

Example (from schema)

{
  "detail": [
    {
      "loc": [
        "string",
        0
      ],
      "msg": "string",
      "type": "string"
    }
  ]
}

Loading...