lightly-ai / lightly-ai/lightly
Implement: Stochastic positional embeddings improve masked image modeling
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 3.8k
- Forks
- 367
- Avg merge
- 3d 22h
- Merged PRs (30d)
- 5
Description
From this paper: https://arxiv.org/pdf/2308.00566
AFAIK we could use that in combination with many other building blocks. Could be interesting to see. E.g. with LeJEPA :)
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
Read the linked paper first, then inspect the existing embedding and LeJEPA-related entry points in the Python/PyTorch library. The issue names no files, tests, integration plan, or acceptance criteria, so define an evaluation against the current masked image modeling behavior before implementation can be considered done.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python, pytorch
- Domain
- computer-vision, machine-learning
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100