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
Commits on main
test(packaging): skip repo-only checks the sdist does not ship
e69ab21 Nithilan Kumaran committed 1 month ago
perf(router): memoize question schema serialization in predict_batch
c907368 Hamidreza committed 1 month ago
test(evals): pin the subclass acceptance and the scalar exactness it keeps
912a4f8 Aashish committed 1 month ago
fix(evals): a JSON container subclass encodes as the mapping it wraps
db74aae Aashish committed 1 month ago
laya-go: delta-review fixes for digests and falsy temperature maps
e71405d Adeshina Lasisi committed 1 month ago
laya-go: review fixes for digests and language temperatures
d653e44 Adeshina Lasisi committed 1 month ago
test(calibrate): a refusal in an acceptance arm is a FAIL line, not a traceback
24cd083 Aashish committed 1 month ago
fix(calibrate): a seed that cannot seed RandomState names itself
a7c97c0 Aashish committed 1 month ago
test(lazy-import): pin laya.mcp's documented no-extra import surface
e3cd18f Aashish committed 1 month ago
go lane: floor 1561 -> 1588 (PR7 tests); drop unused const
1c1c00a Adeshina Lasisi committed 1 month ago