MovieChat: From Dense Token to Sparse Memory for Long Video Understanding

Benchmark Model Rank Results
question-answering-on-next-qa-open-endedMovieChat#5Accuracy: 49.9Confidence Score: 2.7
video-based-generative-performance-1MovieChat#12gpt-score: 2.76
video-based-generative-performance-2MovieChat#13gpt-score: 2.42
video-based-generative-performance-3MovieChat#13gpt-score: 3.01
video-based-generative-performance-4MovieChat#9gpt-score: 2.93
video-based-generative-performance-5MovieChat#13gpt-score: 2.24
video-question-answering-on-activitynet-qaMovieChat#11Accuracy: 45.7Confidence score: 3.1
video-question-answering-on-ovbenchMovieChat (7B)#11AVG: 30.9
zeroshot-video-question-answer-on-activitynetMovieChat#21Accuracy: 45.7Confidence Score: 3.1
zeroshot-video-question-answer-on-msrvtt-qaMovieChat#23Accuracy: 52.7Confidence Score: 2.6
zeroshot-video-question-answer-on-msvd-qaMovieChat#11Accuracy: 75.2Confidence Score: 2.9