3rd Place Solution for Google Universal Image Embedding

October 14, 2022 · Declared Dead · 🏛 arXiv.org

🦴 CAUSE OF DEATH: Skeleton Repo
Boilerplate only, no real code

Repo contents: README.md, model.zip

Authors Nobuaki Aoki, Yasumasa Namba arXiv ID 2210.09296 Category cs.CV: Computer Vision Citations 1 Venue arXiv.org Repository https://github.com/YasumasaNamba/google-universal-image-embedding ⭐ 2 Last Checked 1 month ago
Abstract
This paper presents the 3rd place solution to the Google Universal Image Embedding Competition on Kaggle. We use ViT-H/14 from OpenCLIP for the backbone of ArcFace, and trained in 2 stage. 1st stage is done with freezed backbone, and 2nd stage is whole model training. We achieve 0.692 mean Precision @5 on private leaderboard. Code available at https://github.com/YasumasaNamba/google-universal-image-embedding
Community shame:
Not yet rated
📄 View on arXiv 🌐 View on ar5iv 📑 PDF 💻 Repository 🎉 Report Code Found
Community Contributions

Found the code? Know the venue? Think something is wrong? Let us know!

📜 Similar Papers

In the same crypt — Computer Vision

Died the same way — 🦴 Skeleton Repo