coresets

A research library for pytorch-based neural network pruning, compression, and more.

Coresets via Bilevel Optimization

A library for coreset algorithms, written in JAX for fast execution and GPU support.

Bayesian Coresets Construction with Accelerated Iterative Hard Thresholding (A-IHT).

DataCull is a modular, light-weight data pruning library containing many dataset pruning (coreset selection) algorithm including the official Implementation of the paper, titled, RCAP: Robust, Class-Aware, Probab ilistic Dynamic Dataset Pruning

Geometrical Homogeneous Clustering for Image Data Reduction. An algorithm to reduce large image datasets maintaining similar accuracy.

BICO is a fast streaming algorithm to compute coresets for the k-means problem on very large sets of points.

This is the accompanying code repository for the AISTATS 2022 publication p-Generalized Probit Regression and Scalable Maximum Likelihood Estimation via Sketching and Coresets by Alexander Munteanu, Simon Omlor and Christian Peters.

[ECCV 2022] The official repository of ''$\ell_\infty$-Robustness and Beyond: Unleashing Efficient Adversarial Training''.

Official implementation of "Data-Efficient Training of CNNs and Transformers with Coresets: A Stability Perspective"

Speeding up clustering algorithms using Sampling techniques (Lightweight Coresets)

Efficient image segmentation using KMeans clustering with coreset sampling to speed up processing while preserving segmentation quality.

MECO: Multi-objective Evolutionary Compression

Docs: https://erasus.readthedocs.io/en/latest/ Forget data from any foundation model without retraining. Erasus surgically removes concepts, behaviors, or training samples from LLMs, VLMs, and Diffusion models using coreset selection. 90% less compute, certified removal, multimodal support.

Accelerating SHAP feature attributions using optimal background coresets for out-of-sample robust Explainable AI (XAI).

Implementation of three variants of the Frank-Wolfe method for solving the Minimum Enclosing Ball problem, and application to anomaly detection.

An interactive tool for generating compact, representative coresets from images, enabling faster, memory-efficient processing for tasks like segmentation and clustering with minimal accuracy loss.