DMLC is a community hosting and developing portable, scalable and reliable open source libraries for distributed machine learning. All projects are licenced under Apache 2.0. Contributors come from leading universities and companies. We are warmly welcome new contributors and projects to join it.
Wormhole is a place where DMLC projects works together to provide scalable and reliable machine learning toolkits that can run on various platforms such as Hadoop Yarn, Amazon EC2, MPI, etc.
An optimized general purpose gradient boosting library, including Generalized Linear Model (GLM) and Gradient Boosted Decision Trees (GBDT). XGBoost is parallelized, and also be distributed and scale to Terascale data
A fast, concise, distributed deep learning framework. It provides ready-to-use configure files and trained models for Alexnet, Google Inception network, and others.
High-performance data I/O libraries for various filesystems including local disks, HDFS, Amazon S3. It also provides job launchers for Yarn, MPI, ...
A light weight implementation of the parameter server framework. It provides asynchronous key-value push and pull, communication-efficient data synchronization and flexible data consistency model.
Rabit is a light weight library that provides a fault tolerant interface of Allreduce and Broadcast. The goal of rabit is to support portable , scalable and reliable distributed machine learning programs.
Ph.D. Student at University of Washington
Ph.D. Student at Carnegie Mellon University
Ph.D. Student at NYU
Master's student at Stanford University