各个论文用到的主要数据集
Dataset | Original task | Examples | Sentences |
---|---|---|---|
MSR-VTT | video caption | 10000(20 classes) | ~20 |
KTH | action recognition | 2391 | - |
MSVD | video classification | 1970 | 40 |
Kinetic | video classification | ~500,000(600 classes) | - |
UCF-101 | video classification | 13,320(101 classes)- | |
VaTEX | video caption | ~ 41250 | 10 English+10 Chinese |
MSR-VTT
- 论文 :MSR-VTT: A Large Video Description Dataset for Bridging Video and Language
- 官方下载:ACM Multimedia 2016 Microsoft Research - Video to Text (MSR-VTT) Challenge
其他下载途径:
1.需要翻墙 mediafire
2.hyperAI(https://hyper