microsoft / microsoft/winml-cli
[Tracking] Generate and validate op_check_results for 15 ov_cpu patterns (match/target)
Open
Nobody has claimed this yet.
enhancement
P2
static-analyzer
triaged
- Dominant language
- Python
- Stars
- 40
- Forks
- 11
- Avg merge
- 1d 8h
- Merged PRs (30d)
- 50
Description
Goal
Generate and validate op_check_results for 15 patterns where check_for_eps includes ov_cpu.
Scope (15 patterns)
- Conv2DInplaceLinear2DPattern
- Conv2DInplaceLinear3DPattern
- Conv2DInplaceLinear4DPattern
- ExpandedAttentionPattern
- Gelu1Pattern
- Gelu2Pattern
- Gelu3Pattern
- Gelu4Pattern
- LayerNormalizationMulPattern
- LayerNormalizationPowPattern
- MatMulAddPattern
- ReshapeGemmReshapePattern
- ReshapeTransposeReshapeLowDimPattern
- ReshapeTransposeReshapeOverlyHighDimPattern
- TransposedSingleLayerNormalizationPattern
Validation
- Multi-run stability (run >=3 times and compare case count/case_index/outcomes).
- Compare with previous results (Match/New/Delete/Duplicate deltas).
- Compare with analyze results from pattern split into nodes.
Potential issues to confirm
- Non-determinism across repeated generation.
- Unexpected diffs versus previous baseline.
- Inconsistency between pattern-level checks and node-level analyze outputs.
Definition of done
- All 15 patterns generated.
- Stability verification complete.
- Previous-result comparison complete.
- Pattern-vs-node analyze comparison complete.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Assessment
This issue has not been assessed yet.