Two error

[tetrapod0]

1. 'Upsample' object has no attribute 'recompute_scale_factor'
   Edit torch==1.10.0 in requirements.txt

2. ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (4,) + inhomogeneous part.
   For CPU polygon_detect.py,
   In line numbers 813, 815 utils/general.py,
   Edit boxes1[i, :].view(4,2) -> boxes1[i, :].view(4,2).numpy()
   Edit boxes2[j, :].view(4,2) -> boxes2[j, :].view(4,2).numpy()

[lemon1022]

@TDAT97 Is the calculation of P and R slow when you train?