OpenGVLab / VideoMamba

VideoMamba: State Space Model for Efficient Video Understanding
https://arxiv.org/abs/2403.06977
Apache License 2.0
660 stars 47 forks source link

关于Different scan methods #63

Open Lil-CC opened 3 days ago

Lil-CC commented 3 days ago

想问一下代码中哪里体现的是spatial first bidirectional扫描呢?还有可以选择其他扫描方式的实验吗?

Andy1621 commented 2 days ago

不同的扫描方式由tensor的顺序决定,默认的顺序即为spatial-first