Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.
Python calibrationclassificationdecision-modelhuggingfacejevmodernbertmultilingualnlppythonpytorchroutingtyped-decisionszero-shot
docs(dotnet): Laya.Benchmark is in Laya.slnx, so a solution build does compile it
Adeshina Lasisib31abad51fff
diff --git a/laya-dotnet/samples/Laya.Benchmark/README.md b/laya-dotnet/samples/Laya.Benchmark/README.mdindex 2202e3f2..a3c57c6c 100644--- a/laya-dotnet/samples/Laya.Benchmark/README.md+++ b/laya-dotnet/samples/Laya.Benchmark/README.md@@ -118,6 +118,6 @@ Python runs the PyTorch model directly on the CPU. .NET runs the exported ONNX g## Notes-- **Where the project sits:** it isn't part of `Laya.slnx`, so `dotnet build` / `dotnet test` on the solution skip it. It's a measurement tool, not a shipped sample.+- **Where the project sits:** it is in `Laya.slnx` under `/samples/`, so `dotnet build Laya.slnx` compiles it and a solution build breaks on a compile error here. It holds no tests, so `dotnet test --solution Laya.slnx` runs nothing from it. It's a measurement tool, not a shipped sample.- **Running more than one:** run one checkpoint at a time. The ModernBERT-large checkpoints (english, typed-decisions) are about 1.7 GB each on disk.- **When to measure:** cold latency depends on disk and OS file cache. Run twice and report the second cold number if you want warm file-cache timings.