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(http-api): routing.repo is repo or repo/subfolder, not a public Hugging Face id
Adeshina Lasisi1fdf0d813a8d
diff --git a/docs/http-api.md b/docs/http-api.mdindex d18f42df..639b7214 100644--- a/docs/http-api.md+++ b/docs/http-api.md@@ -226,7 +226,7 @@ caller sizing states by character count cannot see the cut anywhere else in the| `routing` key | meaning ||---|---|| `model` | the checkpoint that answered: `english`, `multilingual` or `typed-decisions` |-| `repo` | its public Hugging Face id |+| `repo` | where it was loaded from, as `repo` or `repo/subfolder`. The default server loads all three from the `convaiinnovations/laya` bundle, so `multilingual` and `typed-decisions` report `convaiinnovations/laya/multilingual` and `convaiinnovations/laya/typed-decisions` -- a readable id, not a Hub repo id to fetch with || `reason` | the sentence for the choice, naming the evidence it acted on || `detection` | `laya.lang.analyse()` on the state -- `script`, `script_profile`, `language`, `is_english`, `language_undecided`, `diacritic_rate`, `non_latin_fraction`, `mixed_segment` -- or `null` when the route decided before reading the text || `workflow` | the typed-decisions workflow the question ids match, or `null` |