- Overview
- Design
- Programming Guide
- Deep Learning Architexture
- Quick Start
- Algorithm
- Deployment
- Community
- FAQ
- Support
- Papers
- Presentation
Angel是一个基于参数服务器(Parameter Server)理念开发的高性能分布式机器学习平台,它基于腾讯内部的海量数据进行了反复的调优,并具有广泛的适用性和稳定性,模型维度越高,优势越明显。 Angel由腾讯和北京大学联合开发,兼顾了工业界的高可用性和学术界的创新性。
Angel的核心设计理念围绕模型。它将高维度的大模型合理切分到多个参数服务器节点,并通过高效的模型更新接口和运算函数,以及灵活的同步协议,轻松实现各种高效的机器学习算法。
Angel基于Java和Scala开发,能在社区的Yarn上直接调度运行,并基于PS Service,支持Spark on Angel,集成了部分图计算和深度学习算法。
欢迎对机器学习有兴趣的同仁一起贡献代码,提交Issues或者Pull Requests。请先查阅: Angel Contribution Guide
Overview
- 架构设计
- 代码结构
- 设计理念
- Spark on Angel
Design
- 模型格式
- 模型切分(modelPartitioner)
- 异步控制(syncController)
- 定制函数(psFunc)
- 核心接口
- 周边辅助
Programming Guide
- Angel编程手册
- Spark on Angel编程手册
Deep Learning Architexture
- Angel中的计算图
- Angel中的层
- Angel中优化器
- Angel中的损失函数
- Angel中的传输函数
- Angel中的学习率Decay
Quick Start
- Angel入门
- Spark on Angel入门
- Angel Json配置
Algorithm
- Angel or Spark On Angel?
- Algorithm Parameter Description
- Angel
- Traditional Machine Learning Methods
- Logistic Regression(LR)
- Support Vector Machine(SVM)
- Factorization Machine(FM)
- Linear Regression
- Robust Regression
- Softmax Regression
- KMeans
- GBDT
- LDA* (WrapLDA)
- Deep Learning Methods
- Deep Neural Network(DNN)
- Mix Logistic Regression(MLR)
- Deep And Wide(DAW)
- Deep Factorization Machine(DeepFM)
- Neural Factorization Machine(NFM)
- Product Neural Network(PNN)
- Traditional Machine Learning Methods
- Spark on Angel
- FTRL
- Logistic Regression(LR)
- Word2Vec
- LINE
- KCORE
- Louvain
- FTRLFM
- GBDT
Deployment
- 下载和编译
- 本地运行
- Yarn运行
- 系统配置
- 资源配置指南
- 使用OpenBlas给算法加速
Community
- Mailing list: angel-tsc@lists.deeplearningfoundation.org
- Angel homepage in Linux FD: https://lists.deeplearningfoundation.org/g/angel-main
- TSC members & Committers
- Contributing to Angel
- Roadmap
FAQ
- 工程类问题
- 算法类问题
Support
- QQ群:20171688
Papers
- Lele Yu, Bin Cui, Ce Zhang, Yingxia Shao. LDA*: A Robust and Large-scale Topic Modeling System. VLDB, 2017
- Jiawei Jiang, Bin Cui, Ce Zhang, Lele Yu. Heterogeneity-aware Distributed Parameter Servers. SIGMOD, 2017
- Jie Jiang, Lele Yu, Jiawei Jiang, Yuhong Liu and Bin Cui. Angel: a new large-scale machine learning system. National Science Review (NSR), 2017
- Jie Jiang, Jiawei Jiang, Bin Cui and Ce Zhang. TencentBoost: A Gradient Boosting Tree System with Parameter Server. ICDE, 2017
- Jiawei Jiang, Bin Cui, Ce Zhang and Fangcheng Fu. DimBoost: Boosting Gradient Boosting Decision Tree to Higher Dimensions. SIGMOD, 2018.
Presentation
Angel: A Machine Learning Framework for High Dimensionality. Strata China, 2017
方圆并济:基于 Spark on Angel 的高性能机器学习. QCon ShangHai China, 2017
基于Angel和Spark Streaming的高维度Online Learning. GIAC China, 2017