Revisit ONNX-specific workarounds
Open
Nobody has claimed this yet.
enhancement
module: models
module: onnx
- Dominant language
- Python
- Stars
- 17.9k
- Forks
- 7.3k
- Avg merge
- 1d 15h
- Merged PRs (30d)
- 13
Description
🚀 Feature
As pointed out by @masahi in https://github.com/pytorch/vision/issues/3221#issuecomment-754924562, torchvision currently contain a few workarounds in order to support ONNX, many of which have been fixed upstream and could be cleaned up.
This issue is to track those potential improvements, so that we can have as little ONNX-specific codepaths as possible.
cc @neginraoof
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 by reading the referenced discussion in issue #3221, then audit torchvision for the ONNX-specific workarounds mentioned there. The issue names no files or tests; done means removing workarounds that are fixed upstream while preserving ONNX support.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- computer-vision, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100