search

leejoonhun / daf

Unofficial PyTorch implementation of Domain Adaptation Forecaster
33 stars 7 forks source link

model's performance #14

Open lear-ner97 opened 9 months ago

lear-ner97 commented 9 months ago

Hi,

Please I'm wondering if you could achieve the same results as shown in the paper after training the model? Because after I trained the model using this code the performance seems far less than the performance showed in the paper. I have attached two figures that show the generator's loss function and the discriminator's loss function. You can see that the generator's loss function does not converge to 0. And the discriminator's loss function is increasing to 0. I'm confused how to interpret this training behaviour. classification_loss generator_loss

Best, Sami

wzhangck commented 2 months ago

hello, Have you found the source code of the paper?