reference-model-inference-images-embeddings.md

title	titleSuffix	description	manager	ms.service	ms.subservice	ms.topic	ms.date	ms.reviewer	ms.author	author	ms.custom
Azure AI Model Inference Image Embeddings	Azure Machine Learning	Reference for Azure AI Model Inference Image Embeddings API	scottpolly	machine-learning	inferencing	conceptual	05/03/2024	None	mopeakande	msakande	build-2024

Reference: Image Embeddings | Azure Machine Learning

Creates an embedding vector representing the input image and text pair.

POST /images/embeddings?api-version=2024-04-01-preview

URI Parameters

Name	In	Required	Type	Description
api-version	query	True	string	The version of the API in the format "YYYY-MM-DD" or "YYYY-MM-DD-preview".

Request Header

Name	Required	Type	Description
extra-parameters		string	The behavior of the API when extra parameters are indicated in the payload. Using pass-through makes the API to pass the parameter to the underlying model. Use this value when you want to pass parameters that you know the underlying model can support. Using ignore makes the API to drop any unsupported parameter. Use this value when you need to use the same payload across different models, but one of the extra parameters may make a model to error out if not supported. Using error makes the API to reject any extra parameter in the payload. Only parameters specified in this API can be indicated, or a 400 error is returned.
azureml-model-deployment		string	Name of the deployment you want to route the request to. Supported for endpoints that support multiple deployments.

Request Body

Name	Required	Type	Description
input	True	EmbeddingInput[]	Input image to embed. To embed multiple inputs in a single request, pass an array. The input must not exceed the max input tokens for the model.
dimensions		integer	The number of dimensions the resulting output embeddings should have. Passing null causes the model to use its default value. Returns a 422 error if the model doesn't support the value or parameter.
encoding_format		EmbeddingEncodingFormat	The format to return the embeddings in. Either base64, float, int8, uint8, binary, or ubinary. Returns a 422 error if the model doesn't support the value or parameter.

Responses

Name	Type	Description
200 OK	CreateEmbeddingResponse	OK
401 Unauthorized	UnauthorizedError	Access token is missing or invalid Headers x-ms-error-code: string
404 Not Found	NotFoundError	Modality not supported by the model. Check the documentation of the model to see which routes are available. Headers x-ms-error-code: string
422 Unprocessable Entity	UnprocessableContentError	The request contains unprocessable content. The error is returned when the payload indicated is valid according to this specification. However, some of the instructions indicated in the payload are not supported by the underlying model. Use the details section to understand the offending parameter. Headers x-ms-error-code: string
429 Too Many Requests	TooManyRequestsError	You have hit your assigned rate limit and your request need to be paced. Headers x-ms-error-code: string
Other Status Codes	ContentFilterError	Bad request Headers x-ms-error-code: string

Security

Authorization

The token with the Bearer: prefix, e.g. Bearer abcde12345

Type: apiKey
In: header

AADToken

Azure Active Directory OAuth2 authentication

Type: oauth2
Flow: application
Token URL: https://login.microsoftonline.com/common/oauth2/v2.0/token

Examples

Creates an embedding vector representing the input text and image

Sample Request

POST /images/embeddings?api-version=2024-04-01-preview

{
  "input": [
    {
      "text": "A nice picture of a cat",
      "image": "data:image/jpeg;base64,iVBORw0KG..."
    }
  ],
  "encoding_format": "float",
  "dimensions": 1024
}

Sample Response

Status code: 200

{
  "data": [
    {
      "index": 0,
      "object": "embedding",
      "embedding": [
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0,
        0
      ]
    }
  ],
  "object": "list",
  "model": "DINO",
  "usage": {
    "prompt_tokens": 15,
    "total_tokens": 15
  }
}

Definitions

Name	Description
ContentFilterError	The API call fails when the prompt triggers a content filter as configured. Modify the prompt and try again.
CreateEmbeddingResponse
CreateImageEmbeddingRequest
Detail
Embedding	Represents an image generated.
EmbeddingEncodingFormat	The format to return the embeddings in. Either base64, float, int8, uint8, binary, or ubinary. Returns a 422 error if the model doesn't support the value or parameter.
EmbeddingObject	The object type, which is always "embedding".
EmbeddingInput	Represents an image with optional text.
ListObject	The object type, which is always "list".
NotFoundError
TooManyRequestsError
UnauthorizedError
UnprocessableContentError	The request contains unprocessable content. The error is returned when the payload indicated is valid according to this specification. However, some of the instructions indicated in the payload are not supported by the underlying model. Use the details section to understand the offending parameter.
Usage	The usage information for the request.

ContentFilterError

The API call fails when the prompt triggers a content filter as configured. Modify the prompt and try again.

Name	Type	Description
code	string	The error code.
error	string	The error description.
message	string	The error message.
param	string	The parameter that triggered the content filter.
status	integer	The HTTP status code.

CreateEmbeddingResponse

Name	Type	Description
data	Embedding[]	The list of embeddings generated by the model.
model	string	The name of the model used to generate the embedding.
object	ListObject	The object type, which is always "list".
usage	Usage	The usage information for the request.

CreateImageEmbeddingRequest

Name	Type	Default Value	Description
dimensions	integer		The number of dimensions the resulting output embeddings should have. Passing null causes the model to use its default value. Returns a 422 error if the model doesn't support the value or parameter.
encoding_format	EmbeddingEncodingFormat	float	The format to return the embeddings in. Either base64, float, int8, uint8, binary, or ubinary. Returns a 422 error if the model doesn't support the value or parameter.
input	EmbeddingInput[]		Input image to embed. To embed multiple inputs in a single request, pass an array. The input must not exceed the max input tokens for the model.

Detail

Name	Type	Description
loc	string[]	The parameter causing the issue
value	string	The value passed to the parameter causing issues.

Embedding

Represents an image generated.

Name	Type	Description
embedding	number[]	The embedding vector, which is a list of floats. The length of vector depends on the model used.
index	integer	The index of the embedding in the list of embeddings.
object	EmbeddingObject	The object type, which is always "embedding".

EmbeddingEncodingFormat

The format to return the embeddings in. Either base64, float, int8, uint8, binary, or ubinary. Returns a 422 error if the model doesn't support the value or parameter.

Name	Type	Description
base64	string
binary	string
float	string
int8	string
ubinary	string
uint8	string

EmbeddingObject

The object type, which is always "embedding".

Name	Type	Description
embedding	string

EmbeddingInput

Represents an image with optional text.

Name	Type	Description
image	string	The input image encoded in base64 string as a data URL. Example: data:image/{format};base64,{data}.
text	string	Optional. The text input to feed into the model (like DINO, CLIP). Returns a 422 error if the model doesn't support the value or parameter.

ListObject

The object type, which is always "list".

Name	Type	Description
list	string

NotFoundError

Name	Type	Description
error	string	The error description.
message	string	The error message.
status	integer	The HTTP status code.

TooManyRequestsError

Name	Type	Description
error	string	The error description.
message	string	The error message.
status	integer	The HTTP status code.

UnauthorizedError

Name	Type	Description
error	string	The error description.
message	string	The error message.
status	integer	The HTTP status code.

UnprocessableContentError

The request contains unprocessable content. The error is returned when the payload indicated is valid according to this specification. However, some of the instructions indicated in the payload are not supported by the underlying model. Use the details section to understand the offending parameter.

Name	Type	Description
code	string	The error code.
detail	Detail
error	string	The error description.
message	string	The error message.
status	integer	The HTTP status code.

Usage

The usage information for the request.

Name	Type	Description
prompt_patches	integer	The number of image patches used by the image prompt.
prompt_tokens	integer	The number of tokens used by the prompt.
total_patches	integer	The total number of patches used by the request.
total_tokens	integer	The total number of tokens used by the request.