chamber111/

VPPO-Eval

Paper: Spotlight on Token Perception for Multimodal Reinforcement Learning Code: https://github.com/huaixuheqing/VPPO-RL

	Dataset Details




	
	Dataset Description

This repository contains the official evaluation dataset for our paper, "Spotlight on Token Perception for Multimodal Reinforcement Learning". It is a collection of 8 challenging multimodal reasoning benchmarks, packaged for easy and reproducible evaluation of the VPPO models.… See the full description on the dataset page: https://huggingface.co/datasets/chamber111/VPPO-Eval.