A Flexible and Powerful Parameter Server for large-scale machine learning

Lightweight and Scalable framework that combines mainstream algorithms of Click-Through-Rate prediction based computational DAG, philosophy of Parameter Server and Ring-AllReduce collective communication.

extremely distributed machine learning

自己实现的深度学习训练框架，纯java实现，没有过多的第三方依赖，可分布式训练

OpenEmbedding is an open source framework for Tensorflow distributed training acceleration.

PetPS: Supporting Huge Embedding Models with Tiered Memory

Serverless ML Framework

A fully adaptive, zero-tuning parameter manager that enables efficient distributed machine learning training

WIP. Veloce is a low-code Ray-based parallelization library that makes machine learning computation novel, efficient, and heterogeneous.

Distributed training with Multi-worker & Parameter Server in TensorFlow 2

Distributed Fieldaware Factorization Machines based on Parameter Server

A parameter server implement with MPI.

Machine Learning models for large datasets

python lib for sparse parameter server using rocksdb, written in c++

Serving layer for large machine learning models on Apache Flink

分布式流数据大规模主题模型的实现

a simple machine learning library

ROS utility package for build-time configuration file generation and dumping/restoring contents of ROS parameter server to/from ROS bags.

A lightweight community-aware heterogeneous parameter server paradigm.