Module types 

Wire types for every SIE endpoint.

Re-exports

pub use cluster::AdaptiveBatchingStatus;

pub use cluster::CapacityInfo;

pub use cluster::ClusterStatusMessage;

pub use cluster::ClusterSummary;

pub use cluster::GpuMetrics;

pub use cluster::HealthResponse;

pub use cluster::ModelCapabilities;

pub use cluster::ModelConfig;

pub use cluster::ModelDims;

pub use cluster::ModelInfo;

pub use cluster::ModelStatus;

pub use cluster::ModelSummary;

pub use cluster::ServerInfo;

pub use cluster::StatusMessage;

pub use cluster::WorkerInfo;

pub use cluster::WorkerStatusMessage;

pub use common::DType;

pub use common::ModelState;

pub use common::OutputDType;

pub use common::OutputType;

pub use common::RequestMetadata;

pub use common::RequestUsage;

pub use common::TimingInfo;

pub use encode::Classification;

pub use encode::DetectedObject;

pub use encode::EncodeResult;

pub use encode::Entity;

pub use encode::ExtractItemError;

pub use encode::ExtractResult;

pub use encode::Multivector;

pub use encode::Relation;

pub use encode::ScoreEntry;

pub use encode::ScoreResult;

pub use encode::ScoreUsage;

pub use encode::SparseVector;

pub use estimate::AppliedRate;

pub use estimate::CostEstimate;

pub use estimate::RateIdentity;

pub use generate::ChatChoice;

pub use generate::ChatChunkChoice;

pub use generate::ChatCompletion;

pub use generate::ChatCompletionChunk;

pub use generate::ChatContent;

pub use generate::ChatContentPart;

pub use generate::ChatDelta;

pub use generate::ChatFinishReason;

pub use generate::ChatImageUrl;

pub use generate::ChatMessage;

pub use generate::ChatRole;

pub use generate::ChatUsage;

pub use generate::FinishReason;

pub use generate::GenerateChunk;

pub use generate::GenerateResult;

pub use generate::GenerationUsage;

pub use generate::Grammar;

pub use generate::ResponseInputMessage;

pub use generate::ResponseOutputMessage;

pub use generate::ResponseOutputText;

pub use generate::ResponseResult;

pub use generate::ResponseUsage;

pub use item::AudioInput;

pub use item::BinaryInput;

pub use item::ImageInput;

pub use item::Item;

pub use jobs::ConnectorJobAttempt;

pub use jobs::ConnectorJobCapabilities;

pub use jobs::ConnectorJobCheckpoint;

pub use jobs::ConnectorJobItemOutcomes;

pub use jobs::ConnectorJobOverlapOwner;

pub use jobs::ConnectorJobPlan;

pub use jobs::ConnectorJobPublication;

pub use jobs::ConnectorJobRecovery;

pub use jobs::ConnectorJobRepairWindow;

pub use jobs::ConnectorJobValidation;

pub use jobs::ConnectorPlanOutputShape;

pub use jobs::JobChunk;

pub use jobs::JobExecution;

pub use jobs::JobFieldMap;

pub use jobs::JobPreflight;

pub use jobs::JobResultItem;

pub use jobs::JobResults;

pub use jobs::JobState;

pub use jobs::JobStatus;

pub use resources::AssignedWorkerInfo;

pub use resources::Batch;

pub use resources::BatchList;

pub use resources::BatchRequestCounts;

pub use resources::Connection;

pub use resources::ConnectionCreated;

pub use resources::ConnectionRevoked;

pub use resources::File;

pub use resources::FileDeleted;

pub use resources::FileList;

pub use resources::PoolInfo;

pub use resources::PoolListItem;

pub use resources::PoolSpecResponse;

pub use resources::PoolStatusInfo;

Modules

cluster

Model catalogue, health and capacity payloads.

common

Scalar wire enums and the per-request metadata every endpoint can return.

encode

Results of /v1/encode, /v1/score and /v1/extract.

estimate

Cost estimation for a request that has not been sent yet.

generate

Text generation: the native /v1/generate shape plus the OpenAI-compatible ones.

item

Request items: the multimodal unit every inference endpoint consumes.

jobs

Batch jobs, including the connector-driven ones.

resources

Files, batches, connections and pools.