NVIDIA / NVIDIA/TensorRT

Fused_MHA does not support when seq_len = 1024, Dh==72, causal_mask==false

Open
#4,104 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

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

Description

When I use FMHA_v2, I found it does not support my scenes. So i wonder is there any way to use fmha_v2 except changing model.
Thx a lot.

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

Start by reproducing the FMHA_v2 case described in the issue: seq_len=1024, Dh=72, and causal_mask=false. Read the TensorRT fused multi-head attention support path to determine why this configuration is rejected and whether the failure is expected. Done means the configuration is supported or the issue clearly documents the required limitation and workaround.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.