Negative loss

[shawrby]

I am currently training the bev model with my dataset.
However, after enough steps, the total loss becomes negative as shown below.

After debugging, when calculating uncertainty_loss(ce_uncertainty_loss, center_uncertainty_loss, offset_uncertainty_loss),
weight(model.module.ce_weight, model.module.center_weight, model.module.offset_weight) is negative.

Are there any possible causes?

[aharley]

Negative loss is OK here. For context you can check these:
https://github.com/yaringal/multi-task-learning-example
https://arxiv.org/abs/1705.07115

[shawrby]

Negative loss is OK here. For context you can check these:
https://github.com/yaringal/multi-task-learning-example
https://arxiv.org/abs/1705.07115

@aharley

Thank you so much!