hyperbolic 开源项目推荐（96 个）

alibaba /Curvature-Learning-Framework

Curvlearn, a Tensorflow based non-Euclidean deep learning framework.

zachetienne /nrpytutorial

NRPy+, BlackHoles@Home, SENRv2, and the NRPy+ Jupyter Tutorial: Python-Based Code Generation for Numerical Relativity... and Beyond!

ariya /ask-llm

Interact with any LLM service

vanzytay /WSDM2018_HyperQA

Reference Implementation for WSDM 2018 Paper "Hyperbolic Representation Learning for Fast and Efficient Neural Question Answering"

FranxYao /PoincareProbe

Implementation of ICLR 21 paper: Probing BERT in Hyperbolic Spaces

reddy-lab /GraphZoo

Facilitating learning, using, and designing graph processing pipelines/models systematically.

agerasev /hypertrace

Real-time hyperbolic space ray tracer

prokolyvakis /hyperkg

HyperKG: Hyperbolic Knowledge Graph Embeddings for Knowledge Base Completion

stevenshave /pybindingcurve

Binding curve simulation and experimental data fitting for multi component protein-ligand systems

HyperbolicLabs /rental-operator

Official Kubernetes operator for Hyperbolic on-demand GPUs

paradite /deepseek-r1-speed-benchmark

Code for benchmarking the speed of DeepSeek R1 from different providers' APIs.

LarsDoorenbos /HoroRF

Repository for "Hyperbolic Random Forests" (TMLR)

CoCoSol007 /hyper

A Hyperbolic 3D Rendering Engine

noamaig /hyperbolic_orbifolds

Implementation of the SIGGRAPH Asia 2016 paper, "Hyperbolic Orbifold Tutte Embeddings"

elichienxD /HyperAid

This is the repository for the KDD 2022 paper "HyperAid: Denoising in hyperbolic spaces for tree-fitting and hierarchical clustering".

dslsdzc /LCM

Multi-lattice codebook retrieval cognitive architecture. JAX training with C99 zero-parameter inference engine.

gabmoreira /hyper

Code for Hyperbolic vs Euclidean Embeddings in Few-Shot Learning: Two Sides of the Same Coin, WACV 2024

okkindel /Hyperbolic

Poincare disk model on HTML5 canvas. Engine ready for development. It allows you to draw polygons, rotate them, move them, it also has a built-in tessellation demo.

gimpong /ICCV25-HLFormer

The code for the paper "HLFormer: Enhancing Partially Relevant Video Retrieval with Hyperbolic Learning" (ICCV'25).

kaustuk000 /Meridian

Meridian is a vision-language representation model built directly on top of a standard CLIP (ViT-B/16) backbone. Unlike typical contrastive models that project image and text representations into a flat Euclidean vector space, Meridian maps multimodal features onto both a continuous hyperbolic manifold (Lorentz manifold) and a flat Euclidean space.