Some questions about loss function.

[Nimo-zl]

Thank you for sharing the great work! I have a confusion about the second stage loss function.

In "def goals_2D_per_example_calc_loss()", there has an nll_loss which caculated by "scores" and "[mapping[i]['goals_2D_labels']]", the "scores" is about dense goals, but the "[mapping[i]['goals_2D_labels']]" seems index from sparse goals. Am I misunderstanding? Why does this work?

[GentleSmile]

It is updated after calculating scores of dense goal

DenseTNT/src/modeling/decoder.py

Line 158 in a0e3b8a

mapping[i]['goals_2D_labels'] = np.argmin(utils.get_dis(goals_2D_dense.detach().cpu().numpy(), label[final_idx]))

[Nimo-zl]

OK, thank you for your replay! =^v^=