FREEDOM251

Freedom251 › Deep Learning Software

Horovod

Free plan APILinuxMacSelf-hosted

Visit Horovod Alternatives Pricing

Horovod's own site

What Horovod is

Horovod is a free, open-source framework for distributed deep learning training with TensorFlow, Keras, PyTorch, and Apache MXNet. It can run on-premise or on AWS, Azure, and Databricks, and can use Apache Spark to combine data processing and model training in one pipeline. Its optimizer averages gradients with allreduce or allgather before applying them. Tensor Fusion groups small allreduce operations and overlaps communication with computation, while automated performance tuning can be enabled from the command line. The project says an existing training script can be scaled to hundreds of GPUs with a few lines of Python code. Documentation covers Docker, Kubernetes, Spark, Ray, Singularity, LSF, and Hadoop Yarn; Gloo is included for running without MPI. Installation guidance provides pip commands for CPU and GPU use with NCCL. Linux and macOS installation requires CMake and has compiler requirements for supported frameworks. The repository says Horovod was archived in September 2026 and is read-only.

Who it suits

Horovod may suit machine-learning teams distributing training across GPUs or using Spark alongside training. The archived, read-only project status is important for anyone considering it for new work.

What it does well

  • Supports TensorFlow, Keras, PyTorch, and Apache MXNet.
  • Can run on-premise, cloud, or Databricks.
  • Includes Gloo for use without MPI.
  • Offers automated performance tuning.
  • Free and open source.

What to know first

  • Repository is archived and read-only.
  • Linux and macOS installation requires CMake.
  • Supported frameworks have compiler requirements.

Limits

Horovod is free, with no trial. Its repository is archived and read-only; Linux and macOS installation requires CMake and compiler requirements for supported frameworks.

Plans

PlanBilled and limitsPrice
HorovodOpen source distributed training frameworkFree

As horovod.ai lists them, read 9 Oct 2026. The maker's own page has the final word.

Getting started

  1. Visit https://horovod.ai/ for project information and getting-started instructions.
  2. Choose a supported environment: Linux, macOS, or a self-hosted deployment.
  3. Install with the provided pip instructions for CPU use or GPU use with NCCL.
  4. Use Gloo if you need to run without MPI installed.
  5. Select a deployment route such as on-premise, AWS, Azure, or Databricks.

Verdict

Horovod provides distributed training tools and integrations across several environments, but its repository is archived and read-only. That project status should weigh heavily in any adoption decision.

Questions

Is Horovod free?

Yes. Horovod is listed as a free, open-source project.

Which frameworks does it support?

TensorFlow, Keras, PyTorch, and Apache MXNet.

What platforms can I use?

The listed platforms are API, Linux, macOS, and self-hosted.

Can it run without MPI?

Yes. Gloo is included as an option for running Horovod without MPI installed.

Is Horovod still maintained?

The repository says it was archived in September 2026 and is read-only.

Facts from Horovod's own pages

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

Where Horovod stands on Freedom251

Look at these too