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(evals): name the entry shapes that read as UNKNOWN instead of as a count
Aashisha1d4532e15e8
diff --git a/docs/evals.md b/docs/evals.mdindex d9be6c21..ed729954 100644--- a/docs/evals.md+++ b/docs/evals.md@@ -616,7 +616,8 @@ laya-evals evidence --checkpoint ./my-checkpoint [--report report.json]A checkpoint directory must contain `rl_agent_config.json`. Calibration evidence is derived fromthe persisted `training.laya_train_calibration` block: a legacy checkpoint with no training-metadata is `UNKNOWN`; a question type with zero items is `MISSING`; a type whose entry records+metadata is `UNKNOWN`; an entry whose `items` is not an integer count — a missing key, a string, a+JSON `true` — is `UNKNOWN` rather than a count of one; a question type with zero items is `MISSING`; a type whose entry recordsany upstream `issues` text (not fitted, below `MIN_TYPE_N`/`CALIB_WARN_N`, clamped, unchangedfit) is `INSUFFICIENT`; a clean fit is `PRESENT`. The helper trusts the persisted `issues`written by #933 instead of keeping a second numeric threshold source.