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
Update script references to use laya_finetune
Ang Zhang389e6be8c7e3
diff --git a/notebooks/laya_finetune_typed_decisions_mps.py b/notebooks/laya_finetune_typed_decisions_mps.pyindex 82a2fefa..44033349 100644--- a/notebooks/laya_finetune_typed_decisions_mps.py+++ b/notebooks/laya_finetune_typed_decisions_mps.py@@ -9,11 +9,11 @@ This is a standalone replacement for the Kaggle 2xT4 notebook. It performs:5. checkpoint/model exportRun without arguments:- python 2.py+python notebooks/laya_finetune_typed_decisions_mps.pyUseful examples:- python 2.py --epochs 2 --micro-batch 1 --grad-accum 32- python 2.py --model-dir ./laya_base --items ./train_items.pt+ python notebooks/laya_finetune_typed_decisions_mps.py --epochs 2 --micro-batch 1 --grad-accum 32+ python notebooks/laya_finetune_typed_decisions_mps.py --model-dir ./laya_base --items ./train_items.ptpython 2.py --device cpu"""