Automation with Laya

Validated Core ML ports of all Laya checkpoints. One short multilingual decision in 4.98 ms P50 on M3 Max with ANE FP16, at 2.78x better energy per decision than MLX FP16.

Swift package that reads a form in a running Mac app through the Accessibility API, asks an on-device model what belongs in each field, and types it in. ~1 ms per decision.

FluidInference ported Laya to CoreML with 99.5% of ops running on the ANE and benchmarked it at 3.7 ms per decision on an M5 Pro.

Ggml-based C++ runtime for all three checkpoints. Native tokenization, decision heads and JSON output; optimized FP16/BF16 CUDA on RTX with FP32 accumulation, plus Vulkan.

Runs the typed-decisions checkpoint on Metal Performance Shaders: ~32 ms median and ~2.1 GiB RAM on M5 Pro, plus a slower ~0.74 GiB low-memory mode. Demo and benchmark included.

Convert → quantize → deploy path that turns Laya (and Kev, NanoJev, PlayJev) into an INT8 ONNX service on the official wire protocol. Runtime is onnxruntime, tokenizers, numpy.

Shows Laya's decisions running in WebAssembly to route steps of a workflow, so branching logic works without a hosted model.

Elixir library that downloads the official checkpoint on first load and answers typed choice and noul questions; pick EMLX on Apple Silicon or EXLA for CPU/NVIDIA.

No Python, no torch, no ONNX export. Loads the ungated safetensors directly and runs any of the three checkpoints on CPU, Metal or CUDA.

Unmodified Laya weights, real NES emulation via Stable Retro, text observations from RAM, model-driven jumping. 27.5 s of game time at 16.79 ms median inference.

Dockerized FastAPI service bundling all three checkpoints behind a language router, with timing-safe API keys, OpenAPI docs, triage and moderation presets and bulk inference.

Anton Lobach ported the Laya System 1 engine to Jolt, a Clojure implementation on Chez Scheme. Code in jlt-commons/lev.

A Java worker that runs Laya's choice, score and yes/no decisions inside Camunda 8 workflows, with versioning and safety checks.

Choice, Score and Noul over any System One backend with a confidence policy: if a backend errors or is unsure, the next is tried. Ships a TypeSafe-compatible server.

Deterministic arcade physics, a browser flight recorder and a supervised Laya fine-tune that landed 30 of 30 held-out flights on Apple MLX. 58% less peak memory with 6-bit weights.

Leanest diffs a PR against main, asks one semantic judgment per spec and runs only the tests it cannot rule out. Andrea Bruno added Laya and classifier.dev as backends next to Jev.