v0.3.3

Full Changelog: v0.3.2...v0.3.3

v0.3.2

Full Changelog: v0.3.1...v0.3.2

v0.3.1

Full Changelog: v0.3.0...v0.3.1

v0.3.0

Full Changelog: v0.2.1...v0.3.0

v0.2.1 — structured criteria fix

Patch release. Fixes a crash on criteria values that are not plain strings.

Fixed

noul questions with structured criteria crashed.

agent.predict(state, {"is_phishing": {
    "type": "noul",
    "instructions": "Is this phishing?",
    "criteria": {"true": {"desc": "phishing, scam or fraud"},
                 "false": {"desc": "legitimate"}}}})
# 0.2.0: TypeError: can only concatenate str (not "dict") to str
# 0.2.1: works, criteria render as JSON

choice and score did not crash but stringified dicts as Python reprs — billing: {'desc': 'payments'} went into the prompt instead of billing: {"desc": "payments"}.

A criterion whose value was 0 or False was treated as missing. The choice branch tested if not v, so those fell back to a bare key. Only None and "" mean "no description" now.

Reported and originally fixed by @trocker in #2 — this release takes the render_criterion approach from that PR on its own, since #2 also carries a large amount of unrelated work.

Added

tests/test_criteria.py — 29 regression tests: the reported crash, JSON rather than repr for dict/list/number criteria, 0/False as real values, non-ASCII, unserialisable objects, and that plain-string criteria are untouched. Both suites now gate CI and releases.

No API changes. Safe upgrade from 0.2.0.

v0.2.0 — model routing

Routes each request to the checkpoint best suited to it.

New

Router picks between laya (English), laya-multilingual (100+ languages) and laya-typed-decisions, loading lazily with LRU eviction.

from laya import Router
router = Router()
router.predict({"body": "I was charged twice"}, questions)          # -> laya
router.predict({"body": "मुझसे दो बार शुल्क लिया गया"}, questions)   # -> laya-multilingual
router.predict(state, questions, model="typed-decisions")           # explicit

route() returns the decision and its reasoning without loading anything.

laya.lang — exact Unicode script detection over 22 scripts plus a best-effort Latin language guess. No new dependencies.

Fixed

• laya-multilingual could not be loaded at all. Its extra_special_tokens ships as a list where transformers expects a mapping, raising AttributeError: 'list' object has no attribute 'keys'.
• ModernBERT's reference_compile now disabled — it torch.compiles the encoder at batch sizes where that is a loss, and can hang.
• Dropped a duplicate email_questions import that shadowed the presets one.

Why route

Measured on a T4, identical questions per model:

	laya	laya-multilingual
MASSIVE intent, English	0.783	0.657
MASSIVE intent, 13 others	0.306	0.451
XNLI, 14 non-English	0.521	0.731

Across all 51 MASSIVE languages the English checkpoint macro-averages 0.227 and clears 3× random on only 23 of 51; the multilingual checkpoint reaches 0.366 and clears it on 45. Khmer scores 0.000 accuracy at 0.952 confidence — the model's own confidence gives no warning, so the decision has to happen before the forward pass.

Note

If laya.load() hangs and TensorFlow is installed, run with USE_TF=0. transformers probes for TF at import and its abseil runtime can deadlock model construction.

Full benchmark data: the research branch.