Hi:
When I try to set --fps to a higher or lower value than the default value of 25, I get an error about tensor size:
"The size of tensor a (XX) must match the size of tensor b (60)"
Is there anything else that needs to be adjusted to match the fps? Or is just a case of the training data being at 25fps?
Regards,