llvm / llvm/torch-mlir

fx_importer_tosa: `BertModule` e2e test times out

Open
#4,172 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
C++
Stars
1.9k
Forks
736
Avg merge
5d 22h
Merged PRs (30d)
15

Description

BertModule e2e test timed out because of TOSA's TosaLayerwiseConstantFoldPass. This is because it takes too long for this pass to go through and fold all constant layerwise ops in the BertModule end-to-end. One potential fix is to apply this pass with a model size limit; if a model is too big, we will skip applying this pass to avoid the time-consuming folds.

Contributor guide

No contributing guide indexed for this repository

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 with the fx_importer_tosa BertModule e2e test and the TosaLayerwiseConstantFoldPass described in the issue. Investigate how the pass processes layerwise constant operations and define a model-size limit so the timed-out test completes without the costly folds.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp, pytorch
Domain
compilers, machine-learning
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
42/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.