Question about batch size and training time

[ziqi-zhang]

Hi,

In your paper you point out that batch size is set to 1 because of implementation, I guess it points to the implementation of DDVO algorithm. But I found that when training posenet, batch size is constrained to 1 in SfMKernel. Is that necessary?
assert(frames.size(0) == 1 and frames.dim() == 5)

And when I train DDVO without posenet it took me 3.5 hours to train half an epoch. I wonder is that a long time normal?

Thanks

[ziqi-zhang]

And I have another question about test dataset. You said in the paper you use 28 sequences for testing, but you also said the performance was evaluated on the same 697 KITTI images from the Eigen's test split. I guess the result table of KITTI test set is based on 697 images. So I don't know how did you use the 28 sequences for testing?