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
release: 0.4.4
Nandakishor682ac6719faa
diff --git a/README.md b/README.mdindex 15bbc9dd..489b36a0 100644--- a/README.md+++ b/README.md@@ -94,26 +94,31 @@ The shipped checkpoints work zero-shot, but fine-tuning on decisions from your o**[nandhakishorm.github.io/laya](https://nandhakishorm.github.io/laya/)**: guides for [prediction hooks](https://nandhakishorm.github.io/laya/hooks/), [schema-driven decisions](https://nandhakishorm.github.io/laya/structured/), [Docker](https://nandhakishorm.github.io/laya/docker/) and [LangChain and LangGraph](https://nandhakishorm.github.io/laya/langchain/), plus a full [API reference](https://nandhakishorm.github.io/laya/reference/).-## What's new in 0.4.3--* **The TypeScript SDK accepts the server's own answers again.** `sdk/typescript`'s response- validation rejected the `x_jev_confidence` 0.4.2 added, which had the CI lane red since that- release; the field is now typed, validated in `[0, 1]`, and absent on `noul` (#1072).-* **`laya.jev_confidence(answer)`** is a public export, so the formula behind the served- `x_jev_confidence` can be applied to a direct `predict()` result (#1083, #302).-* **Choices past 100 options over HTTP.** `LAYA_TOURNAMENT_GROUP=N` answers a larger `choice` by- elimination in groups of N, names every option in `probabilities`, and reports the finalists- under `x_tournament`. Off unless set (#1073, #302).-* **`model: null` is accepted by all four MCP tools**, which is what the shared validator always- meant by `auto` (#1105, #1102).-* **Fixes**: the shortlist cache's width guard reads under the lock it documents (#1104); a falsy- non-object `temperature_by_options` is refused instead of silently loading empty (#1109);- `evaluate()` stops forwarding `model=` to a runner whose `predict` predates it (#1112, #1107);- a question id literally called `questions` is no longer read as a wrapper (#1113, #1110).-* **laya-ts** gains `predictTournament` and a cached embedder, matching Python and Java (#1108,- #1117), and its `min_confidence` bucket-key error now names `default` (#1106, #1100).-* **Java test floors** are remeasured on the merged tree: 1168 and 1169, up from 1007 and 1008,- closing 161 tests of coverage the floor could not see (#1090).+## What's new in 0.4.4++* **laya-go**: a pure-Go tokenizer and language detection, standard library only, with its own+ advisory parity lane that re-records every golden from the Python package at each commit and+ fails on a skipped test (#1145).+* **An out-of-range MCP batch control is reported, not dropped.** `min_confidence=-0.5` or+ `hooks_timeout=-1` reached the tool as "no gate" and "no deadline"; both now come back named+ from the validator (#1118).+* **`min_confidence=0.0` clears a stale abstention marker** instead of returning early, which is+ what `flag_low_confidence` always documented (#1119, #910).+* **`laya --sort-by-length` judges its note per checkpoint group.** The reorder switches inside+ each group, so a file routing to two checkpoints could reorder nothing while the old file-wide+ condition held and printed nothing. `Router.predict_batch` takes a `diagnostics` dict that+ reports it (#1114, #1111).+* **`backend="eager"` pushes an existing replacement aside** rather than reporting itself while a+ graph backend still answers every request, and restores it on uninstall (#1131).+* **laya-ts hook timeouts name the hook and the event**, with Python's `TimeoutError` and message+ (#1140).+* **Evaluation harness**: a mistyped `--model` fails as a usage error before any download (#1120),+ an ONNX run records the commit that answered (#1122), a student's abstention is read from the+ gate's own vocabulary (#1134), and an unreadable runner signature answers both guards the same+ way (#1112).+* **Docs**: measured option-budget ceilings for `/v1/systemone`, the `harm_severity` column's unit+ and why it cannot flag at the default threshold, the AOTInductor route now that it packages, the+ TileLang CPU lowering, and `/health`'s nine-key shape under idle unloading.Earlier releases are in the [GitHub releases](https://github.com/NandhaKishorM/laya/releases).diff --git a/laya/__init__.py b/laya/__init__.pyindex 054efcac..b9695d3c 100644--- a/laya/__init__.py+++ b/laya/__init__.py@@ -35,7 +35,7 @@ from .presets import (from .router import DEFAULT_MODELS, RouteDecision, Routerfrom .structured import DecisionResult, decide, decide_batch-__version__ = "0.4.3"+__version__ = "0.4.4"# Routing, language detection and email cleaning are pure Python. The torch-backed names are# resolved lazily so that `import laya` -- and therefore `from laya import Router` ordiff --git a/nix/package.nix b/nix/package.nixindex c9497402..d9bd762d 100644--- a/nix/package.nix+++ b/nix/package.nix@@ -12,7 +12,7 @@ let# side, where tests/test_packaging.py compares all three declarations (this one,# pyproject.toml and laya.__version__) and fails CI when they diverge. Nothing# did, which is how this sat at 0.3.4 while the package reached 0.3.20.- version = "0.4.3";+ version = "0.4.4";src = ../.;format = "setuptools";propagatedBuildInputs = with python3Packages; [diff --git a/pyproject.toml b/pyproject.tomlindex 6a15585d..d9e83230 100644--- a/pyproject.toml+++ b/pyproject.toml@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"[project]name = "laya"-version = "0.4.3"+version = "0.4.4"description = "Fast, non-autoregressive System 1 decision engine with calibrated probabilities"readme = "README.md"requires-python = ">=3.10"