close this message
arXiv smileybones

arXiv Is Hiring a DevOps Engineer

Work on one of the world's most important websites and make an impact on open science.

View Jobs
Skip to main content
Cornell University

arXiv Is Hiring a DevOps Engineer

View Jobs
We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate
arxiv logo > cs > arXiv:2311.05613

Help | Advanced Search

arXiv logo
Cornell University Logo

quick links

  • Login
  • Help Pages
  • About

Computer Science > Computer Vision and Pattern Recognition

arXiv:2311.05613 (cs)
[Submitted on 9 Nov 2023]

Title:Window Attention is Bugged: How not to Interpolate Position Embeddings

Authors:Daniel Bolya, Chaitanya Ryali, Judy Hoffman, Christoph Feichtenhofer
View a PDF of the paper titled Window Attention is Bugged: How not to Interpolate Position Embeddings, by Daniel Bolya and 3 other authors
View PDF
Abstract:Window attention, position embeddings, and high resolution finetuning are core concepts in the modern transformer era of computer vision. However, we find that naively combining these near ubiquitous components can have a detrimental effect on performance. The issue is simple: interpolating position embeddings while using window attention is wrong. We study two state-of-the-art methods that have these three components, namely Hiera and ViTDet, and find that both do indeed suffer from this bug. To fix it, we introduce a simple absolute window position embedding strategy, which solves the bug outright in Hiera and allows us to increase both speed and performance of the model in ViTDet. We finally combine the two to obtain HieraDet, which achieves 61.7 box mAP on COCO, making it state-of-the-art for models that only use ImageNet-1k pretraining. This all stems from what is essentially a 3 line bug fix, which we name "absolute win".
Comments: Preprint. Code release will be coming in the future
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2311.05613 [cs.CV]
  (or arXiv:2311.05613v1 [cs.CV] for this version)
  https://doi.org/10.48550/arXiv.2311.05613
arXiv-issued DOI via DataCite

Submission history

From: Daniel Bolya [view email]
[v1] Thu, 9 Nov 2023 18:59:58 UTC (3,771 KB)
Full-text links:

Access Paper:

    View a PDF of the paper titled Window Attention is Bugged: How not to Interpolate Position Embeddings, by Daniel Bolya and 3 other authors
  • View PDF
  • TeX Source
  • Other Formats
view license
Current browse context:
cs.CV
< prev   |   next >
new | recent | 2023-11
Change to browse by:
cs

References & Citations

  • NASA ADS
  • Google Scholar
  • Semantic Scholar
a export BibTeX citation Loading...

BibTeX formatted citation

×
Data provided by:

Bookmark

BibSonomy logo Reddit logo

Bibliographic and Citation Tools

Bibliographic Explorer (What is the Explorer?)
Connected Papers (What is Connected Papers?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)

Code, Data and Media Associated with this Article

alphaXiv (What is alphaXiv?)
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Hugging Face (What is Huggingface?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)

Demos

Replicate (What is Replicate?)
Hugging Face Spaces (What is Spaces?)
TXYZ.AI (What is TXYZ.AI?)

Recommenders and Search Tools

Influence Flower (What are Influence Flowers?)
CORE Recommender (What is CORE?)
  • Author
  • Venue
  • Institution
  • Topic

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)
  • About
  • Help
  • contact arXivClick here to contact arXiv Contact
  • subscribe to arXiv mailingsClick here to subscribe Subscribe
  • Copyright
  • Privacy Policy
  • Web Accessibility Assistance
  • arXiv Operational Status
    Get status notifications via email or slack