Horovod

Horovod is a framework for distributing training of Deep Learning models across multiple GPUs, which may be running on multiple machines

Logo

../_images/horovod_horovod-small.png

Website

https://horovod.ai

Repository

https://github.com/horovod/horovod

Byline

Distributed training framework for TensorFlow, Keras, PyTorch, and Apache MXNet.

License

Apache 2.0

Project age

4 years 8 months

Backers

Linux Foundation AI & Data (Governed by), Uber (Creator)

Lastest News (2022-06-21)

v0.25.0 Reducescatter for NCCL, MPI and Gloo, AMD GPU XLA Op implementation, Spark Estimator improvements, TensorFlow Data Service … more

Size score (1 to 10, higher is better)

5.75

Trend score (1 to 10, higher is better)

5.75

Education Resources

URL

Resource Type

Description

https://horovod.readthedocs.io/en/stable/summary_include.html

Documentation

Official project documentation.

https://youtu.be/4y0TDK3KoCA

Video

This is a good tutorial video.

Git Commit Statistics

Statistics computed using Git data through March 31, 2022.

Statistic

Lifetime

Last 12 Months

Commits

45,696

5,231

Lines committed

9,373,276

1,267,585

Unique committers

163

44

Core committers

18

10

../_images/horovod_horovod-monthly-commits.png

Similar Projects

Project

Size Score

Trend Score

Byline

Analytics Zoo

5.25

6.75

Distributed Tensorflow, Keras and PyTorch on Apache Spark/Flink & Ray

Ray

8.75

8.75

An open source framework that provides a simple, universal API for building distributed applications. Ray is packaged with RLlib, a scalable reinforcement learning library, and Tune, a scalable hyperparameter tuning library.