def loss_fn(self, out, annot):
tar_vector = Losses.get_tar_vector(annot)
loss_loc = Losses.get_loc_error(out, tar_vector)
loss_wh = Losses.get_w_h_error(out, tar_vector)
loss_conf = Losses.get_confidence_error(out, tar_vector)
loss_cls = Losses.get_class_error(out, tar_vector)
return loss_loc, loss_wh, loss_conf, loss_cls
some_target_for_loc_error = {
(3, 3):
[[0, [1, 1], [torch.Tensor([0.1]).cuda(),
torch.Tensor([0.1]).cuda()]]]
}
some_predictions_for_loc_error = torch.zeros((1, cell_num, cell_num, 30))
for y in range(cell_num):
for x in range(cell_num):
some_predictions_for_loc_error[0, y, x, :] = torch.Tensor([
0.2, 0.2, 0, 0, 0, 0.2, 0.2, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1, 1, 1, 1, 1, 1, 1, 1
])
unittest.TestCase().assertAlmostEqual(
0.01,
Losses.get_loc_error(some_predictions_for_loc_error,
some_target_for_loc_error).cpu().detach().numpy()[0],
2)
some_targets_class_prob_exist = {
(3, 3): [[1, [0, 0], [torch.Tensor([0]).cuda(),
torch.Tensor([0]).cuda()]]]
}
some_predictions_class_prob_exist = torch.zeros((1, cell_num, cell_num, 30))
for y in range(cell_num):
for x in range(cell_num):
if y == 3 and x == 3:
some_predictions_class_prob_exist[0, y, x, :] = torch.Tensor([
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.2, 0.2, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0
])
else: