Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Transformer example with relative positional encoding #74

Open
albertz opened this issue Nov 10, 2021 · 4 comments
Open

Transformer example with relative positional encoding #74

albertz opened this issue Nov 10, 2021 · 4 comments
Milestone
first-release

Comments

@albertz
Copy link
Member

albertz commented Nov 10, 2021 •
edited
Loading

We should have that example.

Based on Transformer #53 and self attention #52.

Maybe similar to Conformer #54.
@albertz albertz added this to the first-release milestone Nov 10, 2021
@albertz albertz mentioned this issue Nov 10, 2021
Open
@albertz
Copy link
Member Author

albertz commented Jan 3, 2022

Some discussion on variants of pos encoding: https://twitter.com/francoisfleuret/status/1477269757883817985

@albertz albertz mentioned this issue May 2, 2022
Closed
@albertz albertz mentioned this issue Oct 17, 2022
Closed
@albertz
Copy link
Member Author

albertz commented Oct 18, 2022

We have it implemented now, see nn.RelPosSelfAttention.

@albertz
Copy link
Member Author

albertz commented Oct 18, 2022

We still maybe want to have an explicit example of Transformer using nn.RelPosSelfAttention.

@albertz
Copy link
Member Author

albertz commented Nov 2, 2022

Also see: https://github.com/rwth-i6/returnn_common/wiki/Relative-positional-encoding

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
first-release
Development

No branches or pull requests
1 participant
@albertz