VideoGPT+: Integrating Image and Video Encoders for Enhanced Video Understanding

Benchmark Model Rank Results
vcgbench-diverse-on-videoinstructVideoGPT+#1mean: 2.47Correctness of Information: 2.46Detail Orientation: 2.73
video-based-generative-performanceVideoGPT+#7mean: 3.28Correctness of Information: 3.27Detail Orientation: 3.18
video-based-generative-performance-1VideoGPT+#6gpt-score: 3.27
video-based-generative-performance-2VideoGPT+#4gpt-score: 3.39
video-based-generative-performance-3VideoGPT+#6gpt-score: 3.74
video-based-generative-performance-4VideoGPT+#3gpt-score: 3.18
video-based-generative-performance-5VideoGPT+#3gpt-score: 2.83
video-question-answering-on-mvbenchVideoGPT+#10Avg.: 58.7
video-question-answering-on-tvbenchVideoGPT+#18Average Accuracy: 41.7
zeroshot-video-question-answer-on-activitynetVideoGPT+#11Accuracy: 50.6Confidence Score: 3.6
zeroshot-video-question-answer-on-msrvtt-qaVideoGPT+#12Accuracy: 60.6Confidence Score: 3.6
zeroshot-video-question-answer-on-msvd-qaVideoGPT+#15Accuracy: 72.4Confidence Score: 3.6
zeroshot-video-question-answer-on-tgif-qaVideoGPT+#7Accuracy: 74.6Confidence Score: 4.1