You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I’m trying to adapt this example to my own tweets data, but I got really confused about the loss function calculation.
The calculation in the code is not identical to the formula in the paper.
The above two pictures show the Loss function formula in paper, but in the code, it only contains the first part of the sum, L^d. There is a dubious calculation of l underlined in the following picture, which I think might be the second part, but it’s not added to the loss before loss.backward().
And I had to add a multiplication of -clambda to the original prior code (right under the red line in the above picture), because the returned value of model.prior() is actually a minus value of the formula (5) without the lambda factor.
I don’t know Chainer, so it’s very hard to decide whether l was calculated correctly as the second part of the sum and could be readily added to the loss before backward(). It's also hard to decide the intention of the fraction variable and whether it should be multiplied to the l part of the sum too. Actually, I tried adding l to loss without multiplying it to fraction but didn’t get the expected result.
I’m trying to adapt this example to my own tweets data, but I got really confused about the loss function calculation.



The calculation in the code is not identical to the formula in the paper.
The above two pictures show the Loss function formula in paper, but in the code, it only contains the first part of the sum,
L^d. There is a dubious calculation oflunderlined in the following picture, which I think might be the second part, but it’s not added to the loss before loss.backward().And I had to add a multiplication of
-clambdato the originalpriorcode (right under the red line in the above picture), because the returned value of model.prior() is actually a minus value of the formula (5) without the lambda factor.I don’t know Chainer, so it’s very hard to decide whether
lwas calculated correctly as the second part of the sum and could be readily added to thelossbeforebackward(). It's also hard to decide the intention of thefractionvariable and whether it should be multiplied to thelpart of the sum too. Actually, I tried addingltolosswithout multiplying it tofractionbut didn’t get the expected result.