onnx: add com.microsoft MultiHeadAttention handler#2291
Merged
Conversation
Standard (bidirectional) multi-head attention over unpacked query/key/value, lowered onto tract Sdpa, with optional present_key/present_value outputs. Bias, attention/padding masks, packed QKV and past KV cache are rejected with clear errors. Validated bit-close against onnxruntime (output + present_key/value). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
czoli1976
force-pushed
the
feature/onnx-mha
branch
from
c975883 to
8636e5e
Compare
kali
approved these changes
May 27, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
MultiHeadAttention: standard (bidirectional) multi-head attention over unpacked query/key/value, lowered onto tract's Sdpa, with optional present_key/present_value outputs. Bias, attention/padding masks, packed QKV and past KV cache are rejected with clear errors. Validated bit-close vs onnxruntime (output ~1e-7, present-KV bit-exact) across self-/cross-attention, num_heads variations, and the f16 compute path; no node-suite regression; clippy+fmt clean. Part of com.microsoft contrib-op coverage for ORT-exported LLMs.