Table of Contents

Namespace Uralstech.UAI.LiteRTLM

Classes

BackendNames
BenchmarkInfo
Conversation
ConversationConfig
ConversationOptionalArgs
DetokenizeResult
EmbeddingEngine
EmbeddingEngineSettings
EmbeddingOptions
EmbeddingResponse
EmbeddingResponses
Engine
EngineSettings
Experimental

WARNING: The methods in this class are EXPERIMENTAL and subject to change or removal without notice. API stability and backward compatibility are not guaranteed.

InputData
JsonResponse
LiteRTLMErrorReporter

Thread-local error reporting utilities for LiteRT LM operations.

LiteRTLMNativeHandle
LiteRTLMNativeHandleFactory

Utility methods to manually construct wrappers from pointers for types which don't have a native _create method. Use at your own risk.

LiteRTLMNativeLogging
ModelInfo
NoRepeatNgramConfig
RepetitionPenaltyConfig
Responses
SamplerParams
Session
SessionConfig
SessionDebugInfo

WARNING: This class is EXPERIMENTAL and subject to change or removal without notice. API stability and backward compatibility are not guaranteed.

StreamChunk
SuppressTokensConfig
ThinkingConfig
TokenUnion
TokenUnions
TokenizeResult

Structs

BenchmarkInfo.Turn

Benchmarks for a prefill or decode turn.

Responses.Candidate

Enums

ActivationDataType

Represents the activation data type.

BackendType

Hardware backend type.

ConstraintProviderType

Represents the type of constraint provider.

ConstraintType

Represents the type of constraint for constrained decoding.

InputDataType

Represents the type of input data.

InputOverflowStrategy

Strategy for handling inputs longer than the maximum supported signature length.

LogSeverity

Represents the log severity / level.

Modality

Input and output modalities supported by LiteRT-LM models.

ModelType

Model type of the loaded LiteRT-LM file.

NpuBrand

NPU brand options.

SamplerType

Represents the type of sampler.

StatusCode

See the official documentation.

TokenUnionType

Represents the type of a TokenUnion.

Delegates

StreamCallback

Callback for streaming responses.