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
fix(examples): the README row for example 33 listed an option count the example never had
Aashishf0968c52c076
diff --git a/examples/README.md b/examples/README.mdindex dc685825..f5796353 100644--- a/examples/README.md+++ b/examples/README.md@@ -106,7 +106,7 @@ python examples/01_first_call_minimal.py # any Python with torch + trans| `30_custom_schema_design.py` | building a question schema from scratch for your own domain, with the design decisions spelled out. || `31_structured_criteria.py` | criteria as dicts and lists — rendered as compact JSON for the model. || `32_typed_decisions_workflow.py` | the fine-tuned checkpoint's four real workflows, and `auto_task_detection`. |-| `33_high_cardinality_choice.py` | 20, 77 and 120 options: the `head_max_len` option budget, and when to raise it. |+| `33_high_cardinality_choice.py` | 20 and 77 options: the `head_max_len` option budget, and when to raise it. || `34_calibration_temperature.py` | the shipped temperature buckets, and what they do to the logits. |**You can now** write a schema for your own problem and know where its limits are.