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
test(compose): cover compose.modelscope.yaml in the env-passthrough suite
gqchen8417cb057064
diff --git a/tests/test_compose_env.py b/tests/test_compose_env.pyindex 19993fe7..824b4e0e 100644--- a/tests/test_compose_env.py+++ b/tests/test_compose_env.py@@ -41,7 +41,8 @@ FAIL: List[str] = []NOTES: List[str] = []ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))-COMPOSE_FILES = ("compose.yaml", "compose.cuda.yaml", "compose.spark.yaml", "compose.http.yaml")+COMPOSE_FILES = ("compose.yaml", "compose.cuda.yaml", "compose.spark.yaml", "compose.http.yaml",+ "compose.modelscope.yaml")NAME = re.compile(r"\bLAYA_[A-Z0-9_]+\b")# The dtype decision itself: a torch dtype object or the checkpoint's `amp_dtype` field. Prose# that merely names a precision ("MPS fp16 autocast") is not a decision site.