EVA: Efficient Reinforcement Learning for End-to-End Video Agent
Open paper
Benchmark
Model
Rank
Results
video-question-answering-on-videoholmes
EVA (GRPO)
#13
Accuracy: 36.7
video-question-answering-on-mlvu
EVA-GRPO
#40
Accuracy: 68.3
zero-shot-video-question-answer-on-video-mme-w-o-subs
EVA (GRPO)
#47
Accuracy (%): 59.9
Rank counts only results with a code link.