EVERYTHING XIAOMI
Get listed

Horovod

Horovod is a free, open-source framework for distributed deep-learning training with TensorFlow, Keras, PyTorch and Apache MXNet.

Place
#20 of 35
Score
6.0
Horovod's website

What Horovod is

Horovod is a free, open-source framework for distributed deep-learning training with TensorFlow, Keras, PyTorch and Apache MXNet. It is written for Python and can run on Linux, macOS or self-hosted infrastructure. The project describes scaling an existing training script to hundreds of GPUs with a few lines of code. Its distributed optimizer averages gradients through allreduce or allgather before applying them, while Tensor Fusion batches small collective operations and overlaps communication with computation. Automated performance tuning can be enabled from the command line. Horovod can run on-premise or on AWS, Azure and Databricks, and can combine data processing and model training through Apache Spark. Documentation covers Docker, Kubernetes, Spark, Ray, Singularity, LSF and Hadoop Yarn. Gloo is included for installations without MPI. Getting-started instructions provide pip commands for CPU and GPU use with NCCL. Linux and macOS installation requires CMake and suitable compilers. The repository says the project was archived in September 2026 and is read-only.

Good

  • Free open-source distributed training framework
  • Supports TensorFlow, Keras, PyTorch and MXNet
  • Can target hundreds of GPUs
  • Runs with or without MPI
  • Includes automated performance tuning

Watch out

  • Repository archived in September 2026
  • Linux and macOS require CMake
  • Variable deployment setup may require supported compilers

Who it suits

Horovod suits Python teams training TensorFlow, Keras, PyTorch or MXNet models across multiple GPUs. It is also relevant for deployments using Spark or other listed runtime environments.

Verdict

Horovod provides broad distributed-training integrations and GPU support without a license fee. Its archived, read-only repository is an important consideration for new deployments.

Plans and prices

From Horovod's own pricing page, checked 9 Oct 2026.

All about pricing
PlanPriceWhat it gives you
HorovodFreeOpen source distributed training framework

The facts

Each one read from Horovod's own pages, with the date it was checked.

What it does

Autotuning
Horovod includes automated performance tuning that can be enabled with a command-line argument.checked 9 Oct 2026
Collectives
Its distributed optimizer averages gradients using allreduce or allgather before applying them.checked 9 Oct 2026
Deployment
Horovod can be installed on-premise or run on AWS, Azure and Databricks.checked 9 Oct 2026
Installation
The getting-started page provides pip installation commands for CPU use and GPU use with NCCL.checked 9 Oct 2026
MPI alternative
Gloo is included and allows Horovod to run without MPI installed.checked 9 Oct 2026
Open source
Horovod is described as an open source project hosted by the LF AI & Data Foundation.checked 9 Oct 2026
Performance
Tensor Fusion batches small allreduce operations and interleaves communication with computation.checked 9 Oct 2026
Project status
The repository says Horovod was archived in September 2026 due to inactivity and is read-only.checked 9 Oct 2026
Purpose
Horovod is a distributed deep learning training framework for TensorFlow, Keras, PyTorch and Apache MXNet.checked 9 Oct 2026
Requirements
The project README says Linux or macOS installation requires CMake and describes compiler requirements for supported frameworks.checked 9 Oct 2026
Runtime integrations
Documentation describes running Horovod with Docker, Kubernetes, Spark, Ray, Singularity, LSF and Hadoop Yarn.checked 9 Oct 2026
Scale
The project site says an existing training script can be scaled to hundreds of GPUs with a few lines of Python code.checked 9 Oct 2026
Security
The documentation includes a security section for Horovod on Spark; the opened pages make no security certification or compliance claim.checked 9 Oct 2026
Spark
Horovod can run on Apache Spark to combine data processing and model training in one pipeline.checked 9 Oct 2026
Support
The project directs users to documentation, troubleshooting guidance and a ticket submission link for unresolved issues.checked 9 Oct 2026

Alternatives to Horovod

All alternatives
  1. 1

    Paperspace Gradient

    Free planOpen sourceWebAPIExtensionLinux
    8.3From $8 a month
    Visit
  2. 2

    Apache TVM

    Free planOpen sourceAndroidiOSWebAPI
    7.8Free plan
    Visit
  3. 3

    Lightning AI Studios

    Free planWebAPILinuxmacOS
    7.8From $20 a month
    Visit
  4. 4

    ONNX Runtime

    Free planAndroidiOSWebLinux
    7.2Free plan
    Visit
  5. 5

    Ludwig

    Free planAPILinuxmacOSSelf-hosted
    6.7Free plan
    Visit
  6. 6

    MegEngine

    Free planAndroidiOSLinuxmacOS
    6.7Free plan
    Visit