Tensorboard visualization - are the indices correct?

[wpacholarz]

In the snippet marked under the link: https://github.com/zju3dv/LoFTR/blob/c0546af290fe87aaa75f754afe2759e0b00412ea/src/utils/plotting.py#L78C5-L80C67, why are the indices at the end of the line [[1, 0]]?

If I understand the code correctly, assuming a specific b_id, both data['mkpts0_f'], data['mkpts1_f'] and data['scale0'], data['scale1'] have width-related values in the first column and height-related in the second column.

Shouldn't the indices be flipped i.e. isn't the indexing [[1, 0]] redundant? Or have I missed something here? In the current setup the difference is unattainable in visualizations as data['scale0'] and data['scale1'] have almost identical values. It becomes visible when the difference in the scales is significant.

Thank you!