dguo98 / DiffPruning

Parameter Efficient Transfer Learning with Diff Pruning
Apache License 2.0
72 stars 9 forks source link

gradient update #4

Closed Ericmututu closed 2 years ago

Ericmututu commented 2 years ago

Hi, It is a wonderful work. When I tried to understand the source code, I ran into a very confusing problem. The two line codes in line435 and line439 that confuse me. This looks like some kind of chain rule. Is there a clearer explanation?

Thanks in advance.

Ericmututu commented 2 years ago

I understand.

shi-kejian commented 1 year ago

@Ericmututu I struggle at this too -- could you kindly explain here if you've figured it out? Thank you.