deepspeedai/DeepSpeed
DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
From the publisher
README & documentation
Source previewRead the project’s overview, installation instructions and usage examples. The original README is the source of truth.
[](https://github.com/deepspeedai/DeepSpeed/blob/master/LICENSE) [](https://pypi.org/project/deepspeed/) [](https://pepy.tech/project/deepspeed) [](#build-pipeline-status) [](https://www.bestpractices.dev/projects/9530) [](https://twitter.com/intent/follow?screenname=DeepSpeedAI) [](https://twitter.com/DeepSpeedAIJP) [](https://www.zhihu.com/people/deepspeed) [](https://join.slack.com/t/deepspeedworkspace/sharedinvite/zt-3a8pjd8dd-PCj2hMvR4Y2syPwVnjEoww) English · 简体中文 DeepSpeed hosts regular office hours on the last Tuesday of each month at 12:00 America/NewYork to discuss development plans, features, etc. This meeting is public for anyone to join and ask questions. The meeting is hosted on Zoom and can be joined here. More news [2025/04] DeepCompile: Unlocking Compiler Optimization for Distributed Training [2025/03] DeepSpeed AutoTP: Automatic Tensor Parallel Training of Hugging Face models [2024/12] Ulysses-Offload: Democratizing Long Context LLM Training DeepSpeed enabled the world's most powerful language models (at the time of this writing) such as MT-530B and BLOOM. DeepSpeed offers a confluence of system innovations, that has made large scale DL training effective, and efficient, greatly improved ease of use, and redefined the DL training landscape in terms of scale that is possible. These innovations include ZeRO, ZeRO-Infinity, 3D-Parallelism, Ulysses Sequence Parallelism, DeepSpeed-MoE, etc. DeepSpeed was…
Read the full README ↗ · Preview checked 2026-10-31T09:57:22.971Z
Inside the original README — Document outline
- Office Hours
- Latest News
- Extreme Speed and Scale for DL Training
- DeepSpeed Adoption
- Build Pipeline Status
- Installation
- Requirements
- Contributed HW support
- PyPI
- Windows
- Further Reading
- CI funding
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Links from the README
References supplied by the publisher, not independently verified endorsements. Check the destination before downloading files or entering credentials.
- ![License Apache 2.0
- Using Muon Optimizer with DeepSpeed
- System DMA (SDMA) for ZeRO-3: offload collectives off compute units on AMD GPUs for better overlap
- "Building Efficient Large-Scale Model Systems with DeepSpeed: From Open-Source Foundations to Emergi
- Our SuperOffload work received an Honorable Mention for the ASPLOS 2026 Best Paper Award
- DeepSpeed Core API updates: PyTorch-style backward and low-precision master states
- DeepSpeed ZeRO++ powers large-scale distillation training of LLMs for Recommendation Systems at Link
- here
- SuperOffload: Unleashing the Power of Large-Scale LLM Training on Superchips
- Study of ZenFlow and ZeRO offload performance with DeepSpeed CPU core binding
- ZenFlow: Stall-Free Offloading Engine for LLM Training
- Arctic Long Sequence Training (ALST) with DeepSpeed: Scalable And Efficient Training For Multi-Milli
Documentation belongs to its respective authors. Reported project/model license: Apache-2.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
What this repository does
DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
Repository facts
- Owner
- deepspeedai
- Primary language
- Python
- Stars
- 43,278
- Forks
- 5,058
- Open issues + pull requests
- 1,597
- License
- Apache-2.0
- Archived
- No
- Default branch
- master
- Created
- 2020-01-23T18:35:18.000Z
- Last push
- 2026-11-01T11:04:27.000Z
Topics and intended use
Owner-supplied topics: billion-parameters, compression, data-parallelism, deep-learning, gpu, inference, machine-learning, mixture-of-experts, model-parallelism, pipeline-parallelism, pytorch, trillion-parameters, zero
Review the README for scope, installation, examples and limitations. We do not run repository code or certify it.
Evaluate before installing
Review licensing and dependencies, inspect recent commits and unresolved issues, and test in an isolated environment before production use. Stars and forks alone cannot answer these questions.
Source and freshness
Source: GitHub. Metadata observed 2026-11-01T12:19:42.808Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.
Related repositories
EbookFoundation/free-programming-books
donnemartin/system-design-primer