Hi,
I have implemented a sequence to sequence method for my model. However, it doesn’t seem to be modifying my encoder. Any thoughts on why this might be ?
Maybe you forgot to call loss.backward() (or otherwise compute the gradients) or you forgot optimize.step() (or otherwise update the parameters.)
Na. Doing that. Most of the code is the same as the seq2seq tutorial. However there is an off the shelf word embedding being used.
It might help if you posted a link to your code.