Make quantized activation handlers data layout aware #8
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 probably is still rather sketchy, but at least it tries to check the data layout annotation. For now seems to be enough for getting the thresholds of multi-head attention right, IF qonnx properly annotates the 3D layouts.
…-part-map Add V80 to Alveo part_map
15 tasks
|
Looks good, feel free to merge already. |
3 tasks
LinusJungemann
added a commit
that referenced
this pull request
Jun 24, 2025
F backend batch support
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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Outdated suggestions cannot be applied.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.
Cherry-picked and slightly extended commit from Xilinx#901, but unrelated to all the attention stuff. Related to fastmachinelearning/qonnx#143 and the generalized activation function support.
See Xilinx#1183