Update grpo_train.py

jakep-allenai · web-flow · commit 1987422508fb · 2025-11-07T11:46:21.000-08:00
diff --git a/olmocr/train/grpo_train.py b/olmocr/train/grpo_train.py
@@ -1,5 +1,5 @@
 """
-GRPO (Generative Reward-based Policy Optimization) training script for OlmOCR.
+GRPO (Group Relative Policy Optimization) training script for OlmOCR.
 """
 
 import argparse
