[SME][TOPI] Add conv2d NHWC SME fp16->fp32 schedule #17048
Merged
Conversation
This file contains hidden or 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
This commit extends the SME conv2d NHWC schedule to support convolutions with float16 inputs (data and kernel) and a float32 output using the tensor intrinsics added in apache#16981.
This was referenced May 30, 2024
lhutton1
reviewed
Jun 4, 2024
Contributor
lhutton1
left a comment
lhutton1
left a comment
There was a problem hiding this comment.
Apologies for the delay taking a look.. looks great! I just had a couple of questions around handling the weights
Contributor
|
Thanks @Anndrey24 and @lhutton1! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.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 commit extends the SME conv2d NHWC schedule to support convolutions with float16 inputs (data and kernel) and a float32 output using the tensor intrinsics added in #16981.
cc @ekalda @lhutton1