Command Palette
Search for a command to run...
Video Mamba Suite: State Space Model as a Versatile Alternative for Video Understanding
Guo Chen Yifei Huang Jilan Xu Baoqi Pei Zhe Chen Zhiqi Li Jiahao Wang Kunchang Li Tong Lu Limin Wang
Abstract
Understanding videos is one of the fundamental directions in computer vision research, with extensive efforts dedicated to exploring various architectures such as RNN, 3D CNN, and Transformers. The newly proposed architecture of state space model, e.g., Mamba, shows promising traits to extend its success in long sequence modeling to video modeling. To assess whether Mamba can be a viable alternative to Transformers in the video understanding domain, in this work, we conduct a comprehensive set of studies, probing different roles Mamba can play in modeling videos, while investigating diverse tasks where Mamba could exhibit superiority. We categorize Mamba into four roles for modeling videos, deriving a Video Mamba Suite composed of 14 models/modules, and evaluating them on 12 video understanding tasks. Our extensive experiments reveal the strong potential of Mamba on both video-only and video-language tasks while showing promising efficiency-performance trade-offs. We hope this work could provide valuable data points and insights for future research on video understanding. Code is public: https://github.com/OpenGVLab/video-mamba-suite.
Code Repositories
Benchmarks
| Benchmark | Methodology | Metrics |
|---|---|---|
| moment-retrieval-on-charades-sta | video-mamba-suite | R@1 IoU=0.5: 57.18 R@1 IoU=0.7: 36.05 |
| moment-retrieval-on-qvhighlights | video-mamba-suite | R@1 IoU=0.5: 66.65 R@1 IoU=0.7: 52.19 mAP: 45.18 mAP@0.5: 64.37 mAP@0.75: 46.68 |
| temporal-action-localization-on-activitynet | ActionMamba (InternVideo2-6B) | mAP: 42.02 mAP IOU@0.5: 62.43 mAP IOU@0.75: 43.49 mAP IOU@0.95: 10.23 |
| temporal-action-localization-on-fineaction | ActionMamba(InternVideo2-6B) | mAP: 29.04 mAP IOU@0.5: 45.44 mAP IOU@0.75: 28.82 mAP IOU@0.95: 6.79 |
| temporal-action-localization-on-hacs | ActionMamba(InternVideo2-6B) | Average-mAP: 44.56 mAP@0.5: 64.02 mAP@0.75: 45.71 mAP@0.95: 13.34 |
| temporal-action-localization-on-thumos14 | ActionMamba(InternVideo2-6B) | Avg mAP (0.3:0.7): 72.72 mAP IOU@0.3: 86.89 mAP IOU@0.4: 83.09 mAP IOU@0.5: 76.90 mAP IOU@0.6: 65.91 mAP IOU@0.7: 50.82 |
Build AI with AI
From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.