NVIDIA / NVIDIA/TensorRT

No MHA (muti head attention) kernal is called in Tensorrt 10.3 in Orin with Jetpack 6.1

Open
#4,167 7 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

triaged
Dominant language
C++
Stars
13.4k
Forks
2.4k
Avg merge
5d 3h
Merged PRs (30d)
2

Description

Description

Use exact ONNX file attention_ln_opset13.onnx from https://github.com/NVIDIA/TensorRT/issues/3575#issuecomment-1874776406

Attention is like Image

When I use Tensorrt 10.3 with JetPack 6.1, command is trtexec --onnx=attention_ln_opset13.onnx --int8 --saveEngine=default_int8.engine, the nsys as follows doesn't use MHA kernel.

Image

nsys_engine.zip:

  • default_int8.engine
  • default_int8.nsys-rep

Jetpack info

Package: nvidia-jetpack
Source: nvidia-jetpack (6.1)
Version: 6.1+b123
Architecture: arm64
Maintainer: NVIDIA Corporation
Installed-Size: 194
Depends: nvidia-jetpack-runtime (= 6.1+b123), nvidia-jetpack-dev (= 6.1+b123)
Homepage: http://developer.nvidia.com/jetson
Priority: standard
Section: metapackages

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the report with attention_ln_opset13.onnx using the provided trtexec command on TensorRT 10.3 with JetPack 6.1. Inspect default_int8.nsys-rep and default_int8.engine from nsys_engine.zip, then compare the observed execution with the reported absence of an MHA kernel. Done means explaining whether this kernel selection is expected or identifying the cause of the discrepancy.

Written by the indexing model from the issue text.

Assessment

Domain
embedded-iot, machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.