Open qingfeng8-8 opened 2 years ago
Hi, I didn't deal with this use case yet.
But this really depends on the details of the model.
How do the internal activations look like ? Are they images? is every frame processed separately by a model? or are they fused together?
suppose my input is five-dimensional data which includes time dimension(N,C,T,H,W). What do you call the CAM module?Or could open source the code about CAM you have adjusted, thanks a lot.