laya-coreml
Local typed decisions on Apple Core ML with ~5ms inference
Visit laya-coreml ↗
link: rel="ugc noopener"
laya-coreml provides validated ports of Laya decision models to Apple Core ML, enabling on-device inference on Apple's Neural Engine. Achieves ~5ms latency on M3 Max with reproducible performance benchmarks.
Designed for developers building local ML applications on macOS and iOS. Includes energy and speed benchmarks, supporting typed decision workflows without cloud dependencies.
Discussion