MUVERA: Multi-Vector Retrieval via Fixed Dimensional Encoding

Rajesh Jayaram, Laxman Dhulipala, Majid Hadian, Jason Lee, V. Mirrokni
2/3/2026

No official code found. Best implementation: NewBornRustacean/muvera-rs (15 stars, MIT, Rust, last commit 2026-02-18).

Semantic Scholar

Code Implementations(1)

NewBornRustacean/muvera-rsCommunity71%
RustMIT

unofficial implementation of MUVERA: Multi-Vector Retrieval via Fixed Dimensional Encodings

151RustJul 3, 20258 months agoMIT

Similar Papers

InfoGS: Efficient Structure-Aware 3D Gaussians via Lightweight Information Shaping

100% match1 repo32026

Token Statistics Transformer: Linear-Time Attention via Variational Rate Reduction

100% match2026

Feature Directions Matter: Long-Tailed Learning via Rotated Balanced Representation

100% match2026

Referring Image Editing: Object-Level Image Editing via Referring Expressions

100% match2026

InstantPortrait: One-Step Portrait Editing via Diffusion Multi-Objective Distillation

100% match2026

Holodepth: Programmable Depth-Varying Projection via Computer-Generated Holography

100% match2026

Cite this paper

@article{jayaram2026muvera,
  title  = {MUVERA: Multi-Vector Retrieval via Fixed Dimensional Encoding},
  author = {Rajesh Jayaram and Laxman Dhulipala and Majid Hadian and Jason Lee and V. Mirrokni},
  year   = {2026},
  url    = {https://www.semanticscholar.org/paper/430a9f7c7a831fd77ab1af5035df1379b3a49bd7},
  journal = {NEURIPS 2024 2024}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/430a9f7c7a831fd77ab1af5035df1379b3a49bd7.svg)](https://codeofpaper.com/paper/430a9f7c7a831fd77ab1af5035df1379b3a49bd7)

Discussion