Skip to content

docs: add Simplified Chinese translation (README.zh.md) - #8498

Merged
delock merged 4 commits into
deepspeedai:masterfrom
JasonYeYuhe:docs/add-chinese-readme
Sep 27, 2026
Merged

delock merged 4 commits into
deepspeedai:masterfrom
JasonYeYuhe:docs/add-chinese-readme

Conversation

@JasonYeYuhe

Copy link
Copy Markdown
Contributor

Description

This PR provides a comprehensive, high-quality Simplified Chinese translation of the README.md (README.zh.md) to help Chinese-speaking developers, researchers, and engineers explore, adopt, and deploy DeepSpeed in large-scale deep learning training and inference.

Key Changes

  • 1:1 Parity Documentation (README.zh.md): Full translation covering DeepSpeed core innovations (ZeRO, ZeRO-Offload, ZeRO-Infinity, ZeRO++, 3D-Parallelism, Ulysses Sequence Parallelism, DeepSpeed-MoE, SuperOffload, ZenFlow, etc.), hardware accelerator integrations (NVIDIA, AMD, Intel Gaudi/Xeon/XPU, Huawei Ascend NPU, Tecorigin SDAA), installation guides (PyPI, JIT ops, Windows), publications, and video resources.
  • Bidirectional Navigation: Added language switcher between English and Simplified Chinese at the top of both README.md and README.zh.md.
  • Community Maintenance: Includes maintenance notice indicating volunteer maintenance by @JasonYeYuhe.

Checklist

  • Commits are signed off with Developer Certificate of Origin (DCO) (-s)
  • Markdown formatting, tables, image badges, and links have been verified for integrity
  • No code or build system modifications

Signed-off-by: YE <yyyyy.yeyuhe@gmail.com>
Copilot AI lite review requested due to automatic review settings September 13, 2026 04:31

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d77c2de70f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread README.zh.md
@@ -0,0 +1,261 @@
[![License Apache 2.0](https://badgen.net/badge/license/apache2.0/blue)](https://github.com/deepspeedai/DeepSpeed/blob/master/LICENSE)

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Add the required license header

Because README.zh.md is a newly added file, it must begin with the required SPDX and DeepSpeed Team license header; currently the first line is the badge markup, so this file violates the repository's new-file licensing requirement.

AGENTS.md reference: AGENTS.md:L12-L16

Useful? React with 👍 / 👎.

Comment thread README.md
Comment on lines +17 to +18
<p align="center">
<b>English</b> · <a href="README.zh.md">简体中文</a>

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Badge Sign off the non-merge commit

This is a non-merge commit, but its raw commit message has no Signed-off-by: trailer despite the checklist claiming DCO sign-off; add a sign-off using the configured author name and email so the commit meets the repository's DCO requirement.

AGENTS.md reference: AGENTS.md:L8-L8

Useful? React with 👍 / 👎.

Comment thread README.zh.md Outdated
| 英特尔 (Intel) | Intel(R) Gaudi(R) 2 AI 加速器 | hpu | 是 (Yes) | 是 (Yes) |
| 英特尔 (Intel) | Intel(R) 至强(R) 处理器 (Xeon Processors) | cpu | 是 (Yes) | 是 (Yes) |
| 英特尔 (Intel) | Intel(R) 数据中心 GPU Max 系列 (Data Center GPU Max) | xpu | 是 (Yes) | 是 (Yes) |
| 云天励飞/太行 (Tecorigin) | 可扩展数据分析加速器 (SDAA) | sdaa | 是 (Yes) | 否 (No) |

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you double confirm this line whether the name is accurate? Thanks!

Comment thread README.zh.md Outdated
<img src="https://contrib.rocks/image?repo=microsoft/DeepSpeed&r=" width="800px"/>
</a>

## 开发者原创性承诺 (DCO)

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you use “开发者原创性声明” instead?

Comment thread README.zh.md Outdated

* [2026/05] [在 DeepSpeed 中使用 Muon 优化器](https://github.com/deepspeedai/DeepSpeed/blob/master/blogs/muon-optimizer/README.md)

* [2026/05] [针对 ZeRO-3 的系统 DMA (SDMA):将通信算子从 AMD GPU 计算单元卸载,实现更优的通信与计算重叠](https://github.com/deepspeedai/DeepSpeed/blob/master/examples/sdma_allgather/README.md)

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

collective宜翻译成集合通信算子

Comment thread README.zh.md Outdated
</a>

## 开发者原创性承诺 (DCO)
本项目欢迎任何建议与代码贡献。对于绝大多数贡献,您需要同意[开发者原创性承诺 (DCO)](https://wiki.linuxfoundation.org/dco),声明您认可发布于 https://developercertificate.org 的对应条款并拥有授予相关贡献权限的合法权利。

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

的对应条款并拥有 ==> 的对应条款适用于该次特定贡献

pull Bot pushed a commit to AmirulAndalib/DeepSpeed that referenced this pull request Sep 17, 2026
)

## Description

Two kinds of PRs change nothing that CI actually tests, yet currently
run the
full heavy CI (e.g. the python install-smoke matrix and both AWS GPU
workflows):

1. **Docs-only PRs touching root markdown** — `README.md`,
`CONTRIBUTING.md`,
... were not covered by the existing `docs/**` and `blogs/**` exclusions
   (example: deepspeedai#8498, which only touches `README.md` / `README.zh.md`).
2. **Auto-generated release bumps** — the post-release PRs only touch
`version.txt` (example: deepspeedai#8555), which has no effect on any tested
behavior.

This adds root-level `*.md` and `version.txt` exclusions to:

- the `dorny/paths-filter` blocks in `python.yml`, `aws-accelerate.yml`,
and
`aws-torch-latest.yml` (these workflows still trigger and report a
`skipped`
status via their existing `check-paths` job, so Required checks stay
satisfied);
- the `paths-ignore` list in `modal-accelerate.yml` (consistent with the
paths
  already ignored there).

After this change, such PRs run just **Formatting** and **DCO**, plus
the
deliberately-kept light `modal-torch-latest` collect-tests job (it is a
Required check, so it must report a status on every PR; see
`.github/workflows/TEST_SELECTION.md`).

`modal-torch-latest.yml` is intentionally left untouched: adding
`paths-ignore`
to a Required workflow means the check never gets reported and the PR
can never
merge.

## What kind of change does this PR introduce?

CI-only change; no product code is affected.

## How to verify

Open or update a PR that only modifies root `README.md` (or only
`version.txt`)
and confirm that:

- Formatting and DCO run;
- `python` / `aws-accelerate` / `aws-torch-latest` show their
`check-paths` job
  followed by skipped heavy jobs;
- `modal-accelerate` does not trigger at all.

---------

Signed-off-by: Ma, Guokai <guokai.ma@intel.com>
Signed-off-by: YE <yyyyy.yeyuhe@gmail.com>
@JasonYeYuhe

Copy link
Copy Markdown
Contributor Author

Hi @delock, thank you very much for the meticulous and insightful review! 🙏

I have addressed all your feedback in the latest commit:

  1. License Header: Added the required SPDX and DeepSpeed Team license header at the top of README.zh.md per AGENTS.md.
  2. SDMA Collective Communication (Line 31): Updated "通信算子" to "集合通信算子".
  3. Hardware Vendor Accuracy (Line 141): Updated Tecorigin's Chinese name to 太初元碁 (Tecorigin).
  4. DCO Section Heading (Line 192): Changed to 开发者原创性声明 (DCO).
  5. DCO Description (Line 193): Refined phrasing to include "适用于该次特定贡献".
  6. DCO Sign-off: All commits are properly signed off with -s (Signed-off-by:).

Please take another look when convenient. Thank you again for your guidance and support!

@delock

delock commented Sep 19, 2026

Copy link
Copy Markdown
Collaborator

Hi @JasonYeYuhe thanks for the followup. Is the Chinese name of SDAA accurate? Can you check how the vendor call their product name? Its okay if we leave only English product name in the table.

Signed-off-by: YE <yyyyy.yeyuhe@gmail.com>
@JasonYeYuhe

Copy link
Copy Markdown
Contributor Author

Hi @delock, thank you for the helpful suggestion! You're completely right — the vendor (Tecorigin / 太初元碁) refers to their accelerator architecture simply as SDAA, and in technical documentation it is much cleaner to retain the hardware product name in English.

I have updated the table row in README.zh.md to retain SDAA in English:
| 太初元碁 (Tecorigin) | SDAA | sdaa | 是 (Yes) | 否 (No) |

The commit has been signed off with -s and pushed. Please let me know if any further adjustments are needed. Thanks again for your time and guidance! 🙏

@delock
delock enabled auto-merge September 27, 2026 04:01
@delock
delock added this pull request to the merge queue Sep 27, 2026
Merged via the queue into deepspeedai:master with commit 5d2e2c4 Sep 27, 2026
13 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants