Hi all,
I am trying to do multi-task learning, like a single network, multiple loss functions. I show a simple graph below, hope you will get it.
| ---> layer3_1 ---> loss
layer1 ---> layer2 --->
| ---> layer3_2 ---> loss
I already read through issue 224 here, but i can't find the fork function or layer, or split function. Could anybody show me how to do this? Thank you
This guide shows how to share layer via functional API and train a network with multiple losses.
Thanks @joelthchao I will look into it.
Most helpful comment
This guide shows how to share layer via functional API and train a network with multiple losses.