timesfm-ts - v1.0.0
    Preparing search index...

    Class TimesFMNodeEngine

    Implements

    • IInferenceEngine
    Index

    Constructors

    Accessors

    Methods

    • Load the ONNX model from disk and create an inference session.

      Parameters

      • modelPath: string

        Filesystem path to the .onnx model file.

      • Optionaloptions: { skipWarmup?: boolean }

        Load-time flags:

        • skipWarmup: When true, the dummy warmup inference is skipped. This is intended for benchmarking where the caller wants to measure the true first-inference (cold-start) latency separately. Production callers should leave this at the default (false).

      Returns Promise<void>

    • Run a single forward pass through the model.

      Parameters

      • inputs: Float32Array<ArrayBufferLike>[]

        Patched & normed input series [batch, numPatches, inputPatchLen]

      • masks: Uint8Array<ArrayBufferLike>[]

        Patch-level mask [batch, numPatches, inputPatchLen]

      Returns Promise<RawModelOutput>

      Raw model outputs before any post-processing.