Sparsity-aware deep learning inference runtime for CPUs

Libraries for applying sparsification recipes to neural networks with a few lines of code, enabling faster and smaller models

Code for CRATE (Coding RAte reduction TransformEr).

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

Complex-valued neural networks for pytorch and Variational Dropout for real and complex layers.

Repository to track the progress in model compression and acceleration

Official implementation of paper "SparseVLM: Visual Token Sparsification for Efficient Vision-Language Model Inference" proposed by Peking University and UC Berkeley.

[CVPR 2023] Efficient Map Sparsification Based on 2D and 3D Discretized Grids

(Unstructured) Weight Pruning via Adaptive Sparsity Loss

TensorFlow implementation of weight and unit pruning and sparsification

Sparsify Your Flux Models

Feather is a module that enables effective sparsification of neural networks during training. This repository accompanies the paper "Feather: An Elegant Solution to Effective DNN Sparsification" (BMVC2023).

Simple Implementation of the CVPR 2024 Paper "JointSQ: Joint Sparsification-Quantization for Distributed Learning"

TensorFlow implementation of weight and unit pruning and sparsification

An implementation and report of the twice Ramanujan graph sparsifiers.

A simple C++14 and CUDA-based header-only library with tools for sparse-machine learning.

The communication efficiency of federated learning is improved by sparsifying the parameters uploaded by the clients.

Sparsely Reconstructed Graphs (SpaRG) for robust and interpretable fMRI analysis using GNNs and VAEs.