Aligned-MTL-UB Efficient version

median-research-group / LibMTL

A PyTorch Library for Multi-Task Learning

MIT License

1.98k stars 185 forks source link

Closed azj-n closed 8 months ago

azj-n commented 8 months ago

Hello, will you be adding the efficient Aligned-MTL-UB version where they use representation instead of gradient calculation? Thank you

Baijiong-Lin commented 8 months ago

Sure. If adding --rep_grad in the training command, it means computing the gradient w.r.t the shared representation instead of shared parameters.

azj-n commented 8 months ago

I see, thank you.