microsoft / microsoft/onnxruntime
[Feature Request] Request to support rank 3 MaxPool on QNN EP
Nobody has claimed this yet.
- Dominant language
- C++
- Stars
- 21.9k
- Forks
- 4.2k
- Avg merge
- 4d 8h
- Merged PRs (30d)
- 179
Description
Describe the feature request
The current implementation of the pool_op_builder.cc in QNN-EP supports only inputs of rank 4. Support is needed to handle rank 3 as well.
Describe scenario use case
If the original ONNX model is having model input dimension as 3, then the rank of input at the MaxPool node through QNN-EP path will also be 3. This throws error at op support validation stage as it is not meeting the requirement of rank to be 4.
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.
Research direction
Start in pool_op_builder.cc and trace the QNN-EP op support validation that currently requires rank-4 inputs. Verify how a rank-3 ONNX model reaches the MaxPool node, then make rank-3 MaxPool support pass validation while preserving rank-4 behavior. Done means the rank-3 scenario no longer fails at op support validation.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- backend, machine-learning
- Issue type
- Feature
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Clearly specified
- Newbie friendliness
- 48/100