Język

InferenceModelFactory Class

Definition

A factory class for creating instances of the models for mocking.

public static class InferenceModelFactory
type InferenceModelFactory = class
Public Class InferenceModelFactory
Inheritance
InferenceModelFactory

Methods

Name Description
EmbeddingInput(String, Nullable<EmbeddingContentKind>, String, Uri, String)

One item to embed, defaulting to text: an optional id and exactly one source, content or url with sasToken.

EmbeddingItem(String, Int32, String)

Initializes a new instance of EmbeddingItem.

EmbeddingOutput(String, Uri, String, DateTimeOffset, Int32, Int64, String, String)

A delivery: link result: a JSON file with the complete response (the same body as an inline response, including data), readable with sasToken until expiresAt.

EmbeddingResult(String, EmbeddingOutputDataKind, IEnumerable<EmbeddingItem>, EmbeddingUsage, IEnumerable<EmbeddingWarning>, EmbeddingOutput)

The result of an embedding operation.

EmbeddingUsage(String, Nullable<Int64>, Nullable<Int64>)

Initializes a new instance of EmbeddingUsage.

EmbeddingWarning(EmbeddingWarningCode, Nullable<Int32>, Int32, Nullable<Int64>, Nullable<Int64>, String)

Something about a successful multimodal embedding response the caller should know.

LatencyResult(Nullable<TimeSpan>, Nullable<TimeSpan>, Nullable<TimeSpan>)

The result of the latency in the semantic reranking operation.

MultimodalDenseAndMultiVectorEmbedding(Int32, String, EmbeddingContentKind, Nullable<Int32>, Nullable<Int32>, Nullable<EmbeddingImageFormat>, Nullable<Int64>, Nullable<Int64>, Nullable<Boolean>, Nullable<Int64>, IEnumerable<Int32>, BinaryData, BinaryData)

The result for one item when vectorKind is dense_multi: a dense vector and one vector per token, from one model run.

MultimodalDenseEmbedding(Int32, String, EmbeddingContentKind, Nullable<Int32>, Nullable<Int32>, Nullable<EmbeddingImageFormat>, Nullable<Int64>, Nullable<Int64>, Nullable<Boolean>, Nullable<Int64>, BinaryData)

The result for one item when vectorKind is dense: one vector.

MultimodalEmbeddingUsage(Nullable<Int64>, Nullable<Int64>, Nullable<Int64>, Int32, Nullable<Int64>)

Usage for a multimodal embedding operation.

MultimodalMultiVectorEmbedding(Int32, String, EmbeddingContentKind, Nullable<Int32>, Nullable<Int32>, Nullable<EmbeddingImageFormat>, Nullable<Int64>, Nullable<Int64>, Nullable<Boolean>, Nullable<Int64>, IEnumerable<Int32>, BinaryData)

The result for one item when vectorKind is multi: one vector per token, for late-interaction (MaxSim) scoring.

SemanticRerankingContent(String, IEnumerable<String>, Nullable<Boolean>, Nullable<Int32>, Nullable<Int32>, Nullable<Boolean>, Nullable<SemanticRerankingDocumentType>, IEnumerable<String>, String, Nullable<Boolean>)

The input for a semantic reranking operation.

SemanticRerankingMetadata(TokenUsageResult, LatencyResult, String, String)

Additional metadata about the semantic reranking operation.

SemanticRerankingResult(IEnumerable<SemanticRerankingScore>, SemanticRerankingMetadata)

The result of a semantic reranking operation.

SemanticRerankingScore(Nullable<Int32>, String, Nullable<Single>, IEnumerable<SentenceScore>)

A relevance score assigned to a submitted document.

SentenceScore(Int32, Single)

A sentence-level relevance score for a submitted document.

TokenUsageResult(Nullable<Int32>)

The result of the token usage in the semantic reranking operation.

Applies to