Język

InferenceModelFactory.MultimodalEmbeddingUsage Method

Definition

Usage for a multimodal embedding operation.

public static Azure.Data.AI.MultimodalEmbeddingUsage MultimodalEmbeddingUsage(long? promptTokens = default, long? totalTokens = default, long? textTokens = default, int imageCount = 0, long? visualTokens = default);
static member MultimodalEmbeddingUsage : Nullable<int64> * Nullable<int64> * Nullable<int64> * int * Nullable<int64> -> Azure.Data.AI.MultimodalEmbeddingUsage
Public Shared Function MultimodalEmbeddingUsage (Optional promptTokens As Nullable(Of Long) = Nothing, Optional totalTokens As Nullable(Of Long) = Nothing, Optional textTokens As Nullable(Of Long) = Nothing, Optional imageCount As Integer = 0, Optional visualTokens As Nullable(Of Long) = Nothing) As MultimodalEmbeddingUsage

Parameters

promptTokens
Nullable<Int64>

The number of input tokens.

totalTokens
Nullable<Int64>

The total number of tokens.

textTokens
Nullable<Int64>

Tokens that aren't image tokens: the query and text items (including the query instruction) and the short prompt the model adds around each image. Equal to promptTokens minus visualTokens, so text and image tokens can be priced separately. Omitted when the split isn't known: promptTokens isn't reported, or the request has images and visualTokens isn't reported.

imageCount
Int32

The number of images embedded.

visualTokens
Nullable<Int64>

The number of tokens the images produced.

Returns

A new MultimodalEmbeddingUsage instance for mocking.

Applies to