New attention layers #44
Labels
documentation
Improvements or additions to documentation
enhancement
New feature or request
good first issue
Good for newcomers
issue list
A list of issues closely related
jax
Related to JAX
numpy
Related to Numpy
pytorch
Related to Pytorch
Stale
tensorflow
Related to Tensorflow
tests
Related to tests
A list of new attention layers to be added
You can pick any of the following frameworks and implement your layers:
![Tensorflow](https://camo.githubusercontent.com/1e7bc687a8253f83347cecc4c2197c84f8abc6e364c0375127e59638cc7be948/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f54656e736f72466c6f7720322d4646364630303f7374796c653d666c6174266c6f676f3d74656e736f72666c6f77266c6f676f436f6c6f723d7768697465)
![Pytorch](https://camo.githubusercontent.com/8ed9731fde828427756f73b623d8730de4703fa1aa946cd139f6075ea879a4c9/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f5079546f7263682d2532334545344332432e7376673f7374796c653d666c6174266c6f676f3d5079546f726368266c6f676f436f6c6f723d7768697465)
![JAX](https://camo.githubusercontent.com/d50a003396fb7d41e1d477d2df757dc9e8ae4e4e4b14b972d8ebbbc141716eeb/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f4a41582d4138314337442e7376673f7374796c653d666c6174266c6f676f3d6e756d7079266c6f676f436f6c6f723d7768697465)
![Numpy](https://camo.githubusercontent.com/dd9afac0882d3b8e7b5dd05f1c29e8a2b330d453c5cd770e4337ca61e4520340/68747470733a2f2f696d672e736869656c64732e696f2f62616467652f6e756d70792d2532333031333234332e7376673f7374796c653d666c6174266c6f676f3d6e756d7079266c6f676f436f6c6f723d7768697465)
To contribute, please:
Open convert to issue
Please note that we use Numpy style for the docstrings and provide examples as much as possible!
Also please try to provide unit tests using Pytest
Whenever you need help with implementation or any other related issues you might be dealing with, please reach out to me via discussions or Discord server community @soran-ghaderi or @sigma1326
The text was updated successfully, but these errors were encountered: