Compilation is failing Back about pytorch_objectdetecttrack HOT 6 OPEN

I was able to work around this by replacing the import of linear_assignment with:

from scipy.optimize import linear_sum_assignment
def linear_assignment(x):
  indices = linear_sum_assignment(x)
  indices = np.asarray(indices)
  return np.transpose(indices)

from pytorch_objectdetecttrack.

Yes， sort.py has the same problem in my side ------TypeError: tuple indices must be integers or slices, not tuple, can you help?

from pytorch_objectdetecttrack.

Am trying to solve it, if I do it I'll post here.

from pytorch_objectdetecttrack.

I resolved my first error by a simple change "if(d not in matched_indices[0]):" i.e. Only access the 0 and 1 element of tuple.
But now I have problem in matching section.

from pytorch_objectdetecttrack.

Getting the same issue on sort.py with videos. The only change I made was altering the sklearn module to scipy in sort.py.

from pytorch_objectdetecttrack.

I was able to work around this by replacing the import of linear_assignment with:

from scipy.optimize import linear_sum_assignment
def linear_assignment(x):
  indices = linear_sum_assignment(x)
  indices = np.asarray(indices)
  return np.transpose(indices)

This also worked great for me. Thank you!

from pytorch_objectdetecttrack.