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): example 33 quoted the shipped temperature, not the applied one
Aashish60bddb9dabd0
diff --git a/examples/33_high_cardinality_choice.py b/examples/33_high_cardinality_choice.pyindex 0175c2ca..fb7537b9 100644--- a/examples/33_high_cardinality_choice.py+++ b/examples/33_high_cardinality_choice.py@@ -145,6 +145,7 @@ for head, ctx in ((512, 1024), (1024, 2048)):budget_report(agent, seventy_seven, ctx, head)ask(agent, seventy_seven, "77 options, raised")-print("\n confidence is 1.000 either way: the `choice:11+` temperature (0.1006) sharpens the")+print("\n confidence is 1.000 either way: the `choice:11+` bucket applies 0.5 -- its shipped")+print(" 0.1006 is below the floor and is clamped into [0.5, 5] at load -- which sharpens the")print(" logits, so it cannot warn you that a 4-token option was misread. Raise the budget or")print(" split the label space; do not trust confidence on large option sets.")