cyx-6 opened a new pull request, #14744: URL: https://github.com/apache/tvm/pull/14744
In this PR, we add the support for the implicit attention pattern matching in BYOC. By adding a rewriting stage before `FuseOpsByPattern` , we are able to match and rewrite the patterns like `matmul - softmax - matmul` to `R.nn.attention` with other operators needed. cc: @masahi @yelite -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
