Use attention dropout during training

#10
No description provided.
Jina AI org

Can you move dropout_p to the constructor?

You mean instead of using self.dropout.p we use something like self.dropout_p that we set in the constructor? We will still need to check self.training in the forward pass.

Markus28 changed pull request status to open
Markus28 changed pull request status to merged

Sign up or log in to comment