embeddingIntraOpNumThreads property

int? embeddingIntraOpNumThreads
final

Maximum number of threads for intra-op parallelism in ONNX runtime.

If threadLevel is set, this value is ignored.

Set this to a small number (e.g., 1 or 2) to reduce CPU usage and heat on mobile devices, at the cost of slower embedding speed. If both are null, defaults to ~50% of available cores.

Implementation

final int? embeddingIntraOpNumThreads;