[Unity][Frontend][NN] Add diffusers style Attention layer #15609
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from Reviewers by @-ing them in a comment.
Generated by tvm-bot |
jwfromm
changed the title
|
cc @junrushao @cyx-6 |
LeshengJin
approved these changes
Aug 24, 2023
added 5 commits
August 25, 2023 17:23
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds support for the
Attentionlayer in the nn module API. This layer mimics the behavior of the Attention layer used in huggingface Diffusers. Under the hood it uses scaled dot product attention. Notably, there are still some missing features. For example I didnt add support for attention masks yet. I also am assuming 3 dimensional inputs so this likely wouldnt be useful for the VAE of stable diffusion. However, it does allow us to represent the all attention layers of the UNET. I'm hoping we can expand on the functionality more if needed in future PRs.