Use fuse/expand_dims for squeeze unsqueeze
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 5k
- Forks
- 1.2k
- Avg merge
- 2d 10h
- Merged PRs (30d)
- 581
Description
🚀 The feature, motivation and pitch
Right now to represent squeeze/unsqueeze (for conv1d) in our graph, we use XNNPACK's static_reshape operator. This has some limitations on dynamism. Let's instead move to the new operators (fuse_dims and expand_dims) in XNNPACK. This should provide better dynamism support around our squeeze/unsqueezes.
Alternatives
No response
Additional context
No response
RFC (Optional)
No response
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
Locate the current conv1d squeeze/unsqueeze handling that emits XNNPACK's static_reshape operator, then read the fuse_dims and expand_dims operator definitions. Done means the graph uses the newer operators and retains better dynamism support around squeeze and unsqueeze operations.
Written by the indexing model from the issue text.
Assessment
- Domain
- embedded-iot, machine-learning
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100