HawkEye: Training Video-Text LLMs for Grounding Text in Videos
Open paper
Benchmark
Model
Rank
Results
moment-retrieval-on-charades-sta
HawkEye 7B (fine-tuned) (extra data)
#42
R@1 IoU=0.5: 58.3
R@1 IoU=0.7: 28.8
R@1 IoU=0.3: 72.5
mIoU: 49.3
video-question-answering-on-mvbench
HawkEye
#82
Avg.: 47.55
Rank counts only results with a code link.