Class InferenceConfig

Note

Construct via the Class AniraWeb factory in most cases: aniraWeb.InferenceConfig(...) rather than new InferenceConfig(...). The factory threads the WASM instance through for you.

class InferenceConfig(wasmInstance, modelDataVector, tensorShapeVector, processingSpec, maxInferenceTime, warmUp, sessionExclusiveProcessor, blockingRatio, numParallelProcessors)

TypeScript wrapper for anira::InferenceConfig Thread-safe C API wrapper

Extends:

  • BaseWrapper

Arguments:

  • wasmInstance (AniraWasmInstance)

  • modelDataVector (PossiblePointer<VectorModelData>)

  • tensorShapeVector (PossiblePointer<VectorTensorShape>)

  • processingSpec (PossiblePointer<ProcessingSpec>)

  • maxInferenceTime (number)

  • warmUp (number)

  • sessionExclusiveProcessor (boolean)

  • blockingRatio (number)

  • numParallelProcessors (number)

destroy()

Free the underlying C++ object. See Lifecycle and Cleanup for when to call this.

equals(other)
Arguments:

  • other (PossiblePointer<InferenceConfig>)

Returns:

boolean

getBlockingRatio()

Mirrors the anira::InferenceConfig::m_blocking_ratio field.

Returns:

number

getInternalModelLatency(tensorIndex=0)

Mirrors anira::InferenceConfig::get_internal_model_latency().

Arguments:

  • tensorIndex (number)

Returns:

number

getMaxInferenceTime()

Mirrors the anira::InferenceConfig::m_max_inference_time field.

Returns:

number

getModelData(backend)

Mirrors anira::InferenceConfig::get_model_data().

Arguments:

  • backend (number)

Returns:

number

getModelPath(backend)

Mirrors anira::InferenceConfig::get_model_path().

Arguments:

  • backend (number)

Returns:

string

getPostprocessOutputChannels(tensorIndex=0)

Mirrors anira::InferenceConfig::get_postprocess_output_channels().

Arguments:

  • tensorIndex (number)

Returns:

number

getPostprocessOutputSize(tensorIndex=0)

Mirrors anira::InferenceConfig::get_postprocess_output_size().

Arguments:

  • tensorIndex (number)

Returns:

number

getPreprocessInputChannels(tensorIndex=0)

Mirrors anira::InferenceConfig::get_preprocess_input_channels().

Arguments:

  • tensorIndex (number)

Returns:

number

getPreprocessInputSize(tensorIndex=0)

Mirrors anira::InferenceConfig::get_preprocess_input_size().

Arguments:

  • tensorIndex (number)

Returns:

number

getTensorInputShape(backend)

Mirrors anira::InferenceConfig::get_tensor_input_shape().

Returns a non-owning TensorShapeList view into the config’s universal (or per-backend) input shape. Do not call .destroy() on it — the underlying storage belongs to the InferenceConfig.

Arguments:

  • backend (number)

Returns:

TensorShapeList

getTensorInputSize(tensorIndex=0)

Mirrors anira::InferenceConfig::get_tensor_input_size().

Arguments:

  • tensorIndex (number)

Returns:

number

getTensorOutputShape(backend)

Mirrors anira::InferenceConfig::get_tensor_output_shape().

Returns a non-owning TensorShapeList view. See getTensorInputShape for ownership notes.

Arguments:

  • backend (number)

Returns:

TensorShapeList

getTensorOutputSize(tensorIndex=0)

Mirrors anira::InferenceConfig::get_tensor_output_size().

Arguments:

  • tensorIndex (number)

Returns:

number

getTensorShape(backend)

Mirrors anira::InferenceConfig::get_tensor_shape().

Returns a non-owning TensorShape view selected for backend (falling back to the universal shape if no backend-specific shape is registered). Do not call .destroy() on it.

Arguments:

  • backend (number)

Returns:

TensorShape

getWarmUp()

Mirrors the anira::InferenceConfig::m_warm_up field.

Returns:

number

isModelBinary(backend)

Mirrors anira::InferenceConfig::is_model_binary().

Arguments:

  • backend (number)

Returns:

boolean

notEquals(other)
Arguments:

  • other (PossiblePointer<InferenceConfig>)

Returns:

boolean

setBlockingRatio(value)

Mirrors the anira::InferenceConfig::m_blocking_ratio field.

Arguments:

  • value (number)

setMaxInferenceTime(value)

Mirrors the anira::InferenceConfig::m_max_inference_time field.

Arguments:

  • value (number)

setWarmUp(value)

Mirrors the anira::InferenceConfig::m_warm_up field.

Arguments:

  • value (number)