Some questions about the forward_features function in the videomamba_pretrain file

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

Some questions about the forward_features function in the videomamba_pretrain file #20

Closed Ly403 closed 3 months ago

Ly403 commented 3 months ago

Hi! Thanks for providing such a remarkable work. But when reproducing your work, I have encountered some problems.

I found that in the line 432 in the videomamba_pretrain.py file, you passed four arguments (including the self argument) into the forward_features function, while in line 368 it seems that the forward_features function accepts only three arguments (including the self argument). Is that a bug, or am I making something wrong?

Many thanks. Looking forward to your reply.

Andy1621 commented 3 months ago

Thanks for your question! It's a bug when I cleaning the code, and I have fixed it~

Ly403 commented 3 months ago

OK. Thanks a lot!