You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
If i've got your model right, it takes 3 images - I1, I2, I3 and gives flow between I2 and I3 (and backward, I3 -> I2). So, question is. How to evaluate your model on full Sintel dataset to get error value if, as i got it, i need to skip first image in every folder? So, opening folder alley_1 and sending:
Frame_1 Frame_2 Frame 3 to the network and getting flow between Frame 2 and Frame 3, calculating differences and going further. But how you managed to calculate those numbers in the table without that Frame_1 -> Frame_2 flow? Or am i missing something? I've seen that you have DDFlow, but, as i understand, it's a different network. Or..?
Thanks in advance for the explanation
The text was updated successfully, but these errors were encountered:
Hello,
If i've got your model right, it takes 3 images - I1, I2, I3 and gives flow between I2 and I3 (and backward, I3 -> I2). So, question is. How to evaluate your model on full Sintel dataset to get error value if, as i got it, i need to skip first image in every folder? So, opening folder alley_1 and sending:
Frame_1 Frame_2 Frame 3 to the network and getting flow between Frame 2 and Frame 3, calculating differences and going further. But how you managed to calculate those numbers in the table without that Frame_1 -> Frame_2 flow? Or am i missing something? I've seen that you have DDFlow, but, as i understand, it's a different network. Or..?
Thanks in advance for the explanation
The text was updated successfully, but these errors were encountered: