Loss linear combination

[gabrer]

Are the coefficients used while summing up the losses fixed? How did you choose them?
In the original paper, they would be alpha_t, where t is the particular task.

Thank you!

[FrankWork]

I think it is a hyper-parameter. I choose one with higher F1 score.

[gabrer]

Ok, thank you!
Still looking for any other approach.