It keeps the JS surface intentionally compact: dense distances, dot products, binary metrics, probability divergences, packed GEMM-like matrix multiplication, symmetric Gram matrices, dtype-tagged low-precision storage, typed views, and runtime capability inspection.
Topics
| Installation | The package targets Node >= 22. |
|---|---|
| Quickstart | |
| Highlights | This SDK is deliberately smaller than Python or Rust. |
| Scalar Types34 | JavaScript has no built-in f16, bf16, fp8, or packed-bit numeric model. |
| Tensors and Views75 | The wrapper hierarchy exists to keep dtype and ownership explicit across addon and WASM boundaries. |
| Dot Products2 | Dot products are separate from distances because dtype tagging and low-precision storage matter more here. |
| Dense Distances3 | Dense distance entrypoints work directly on the standard numeric TypedArray types. |
| Set Similarity3 | Binary metrics operate on packed storage rather than on generic boolean arrays. |
| Probability Metrics2 | These call the underlying SIMD divergence kernels directly. |
| Matrix Kernels18 | |
| Runtime Dispatch10 | |
| Platform Notes | The npm package includes pre-built WASM modules under wasm/, one per SIMD capability and address width: numkong-wasm32-serial, numkong-wasm32-v128, numkong-wasm32-v128relaxed, and numkong-wasm64-v128relaxed. |