How to do modality fusion

Question

yojayc opened this issue 5 years ago · comments

Anyone knows how to implement modality fusion to get fusion result of rgb stream and flow stream in the pytorch version of tsn?

Robinxin123 · Answer 1 · Fri Jun 28 2019 19:31:32 GMT+0800 (China Standard Time)

hello, i have the same question, do you solve it ? thanks

yojayc · Answer 2 · Fri Jun 28 2019 20:13:12 GMT+0800 (China Standard Time)

hello, i have the same question, do you solve it ? thanks

I use the Video-level testing part of the original version of TSN to calculate the final results.

Robinxin123 · Answer 3 · Fri Jun 28 2019 21:02:27 GMT+0800 (China Standard Time)

ah, thanks, i got the same solution in another issue.

yjxiong · Answer 4 · Wed Jul 10 2019 08:02:57 GMT+0800 (China Standard Time)

Yes. You can use the Caffe TSN's Python scripts for modaity fusion tests.

This is an experimental project. We have released a full fledged toolbox for Pytorch at MMAction.