请问，这两个损失值可以分开反向传播吗

znxlwm / pytorch-MNIST-CelebA-GAN-DCGAN

Pytorch implementation of Generative Adversarial Networks (GAN) and Deep Convolutional Generative Adversarial Networks (DCGAN) for MNIST and CelebA datasets

507 stars 145 forks source link

请问，这两个损失值可以分开反向传播吗 #11

Open lijiajun3029 opened 4 years ago

lijiajun3029 commented 4 years ago

https://github.com/znxlwm/pytorch-MNIST-CelebA-GAN-DCGAN/blob/c88438c7807326492757623bf9117ee2eb5de8bf/pytorch_MNIST_DCGAN.py#L210

jizongFox commented 3 years ago

But what is the propose of separating them?

wangqiang9 commented 2 years ago

https://github.com/znxlwm/pytorch-MNIST-CelebA-GAN-DCGAN/blob/c88438c7807326492757623bf9117ee2eb5de8bf/pytorch_MNIST_DCGAN.py#L210

可以