Update README.md

README.md

@@ -21,7 +21,7 @@ license: mit
 
 ## Overview
 
-**Omni-R1-Zero** is trained **without multimodal annotations**. It bootstraps **step-wise visualizations** from **text-only CoT seeds** (e.g., M3CoT), and then follows the same **SFT → RL** recipe as Omni-R1 to learn interleaved multimodal reasoning.
+**Omni-R1-Zero** is trained **without multimodal annotations**. It bootstraps **step-wise visualizations** from **text-only CoT seeds** (e.g., M3CoT), and then follows the same PeSFT+PeRPO recipe as Omni-R1 to learn interleaved multimodal reasoning.
 
 ## Usage