Progressive Video Condensation with MLLM Agent for Long-form Video Understanding
Open paper
Benchmark
Model
Rank
Results
zero-shot-video-question-answer-on-egoschema-fullset
ProVCA
–
Accuracy: 69.3
zero-shot-video-question-answer-on-intentqa
ProVCA
–
Accuracy: 77.7
zero-shot-video-question-answer-on-next-qa
ProVCA
–
Accuracy: 80.5
Rank counts only results with a code link.