HyperAI超神经

Visual Reasoning On Phyre 1B Within

评估指标

AUCCESS

评测结果

各个模型在此基准测试上的表现结果

比较表格
模型名称AUCCESS
forward-prediction-for-physical-reasoning80.0
physical-reasoning-using-dynamics-aware85.2
learning-long-term-visual-dynamics-with85.2
phyre-a-new-benchmark-for-physical-reasoning77.6