[diffusion] docs: consolidate diffusion documentation into docs by qianyue76 · Pull Request #18095 · sgl-project/sglang

qianyue76 · 2026-02-02T05:33:19Z

Overview

Consolidate scattered SGLang-Diffusion documentation into a unified docs/diffusion/ directory with improved organization.

Before

Documentation was scattered across multiple locations:

docs/supported_models/diffusion_models.md (1279 lines)
docs/supported_models/diffusion_language_models.md
python/sglang/multimodal_gen/docs/ (14+ files including cli.md, openai_api.md, cache/, etc.)

After

docs/diffusion/
├── index.md                    # Landing page
├── installation.md             # Installation (merged ROCm)
├── compatibility_matrix.md     # Supported models
├── api/                        # API documentation
│   ├── cli.md
│   └── openai_api.md
├── performance/                # Performance optimization
│   ├── index.md               # Overview (NEW)
│   ├── attention_backends.md
│   ├── profiling.md
│   └── cache/
│       ├── index.md
│       ├── cache_dit.md
│       └── teacache.md
├── support_new_models.md
├── contributing.md
├── environment_variables.md
├── ci_perf.md
└── language_models/
   └── index.md               # LLaDA

Key Changes

Unified Location: All diffusion docs moved to docs/diffusion/
Logical Grouping: api/ for interfaces, performance/ for optimization
Reduced Duplication:
- Merged installation_rocm.md into installation.md
- Removed duplicate env vars table from cache_dit.md
Better Navigation: Added performance/index.md overview page
Backward Compatibility: Internal docs have redirect notices

Files Changed

21 files changed, 130 insertions(+), 65 deletions(-)
7 files renamed/reorganized to new structure
1 file created (performance/index.md)
1 file deleted (installation_rocm.md)

Motivation

Modifications

Accuracy Tests

Benchmarking and Profiling

Checklist

Format your code according to the Format code with pre-commit.
Add unit tests according to the Run and add unit tests.
Update documentation according to Write documentations.
Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.
Follow the SGLang code style guidance.

Review Process

Ping Merge Oncalls to start the PR flow. See the PR Merge Process.
Get approvals from CODEOWNERS and other reviewers.
Trigger CI tests with comments or contact authorized users to do so.
- /tag-run-ci-label, /rerun-failed-ci, /tag-and-rerun-ci
After green CI and required approvals, ask Merge Oncalls to merge.

gemini-code-assist · 2026-02-02T05:33:46Z

Summary of Changes

Hello @qianyue76, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request significantly refactors the SGLang-Diffusion documentation by centralizing all related content into a dedicated docs/diffusion directory. This initiative aims to enhance the overall user experience by providing a more intuitive and organized structure for accessing guides, API references, and model compatibility information. The changes involve creating new documentation files in the designated location, updating the main table of contents, and implementing redirect notices to ensure seamless navigation for existing links.

Highlights

Documentation Consolidation: All SGLang-Diffusion documentation has been moved from various locations into a new, centralized docs/diffusion/ directory for improved organization and discoverability.
New Directory Structure: A new directory structure has been established under docs/diffusion/, including subdirectories for cache/ and language_models/, to logically group related content.
Table of Contents Update: The main docs/index.rst file has been updated to include a new 'SGLang Diffusion' section, while removing the old, scattered diffusion entries from the 'Supported Models' section.
Backward Compatibility: Redirect notices have been added to the original documentation files to ensure users are guided to the new locations, maintaining a smooth transition.
Internal Link Updates: Links within python/sglang/multimodal_gen/README.md and other internal documentation files have been updated to reflect the new documentation paths.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	`/gemini review`	Performs a code review for the current pull request in its current state.
Pull Request Summary	`/gemini summary`	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	`/gemini help`	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

gemini-code-assist

Code Review

This pull request is a significant and well-executed refactoring that consolidates all SGLang-Diffusion documentation into a new, unified docs/diffusion/ directory. This greatly improves the organization and maintainability of the documentation. The changes include creating new documentation files, updating the main index to reflect the new structure, and thoughtfully adding redirection notices to old file locations for backward compatibility. My review focuses on ensuring the new documentation is clear, consistent, and that all links are correct. I've identified a few minor areas for improvement regarding link targets, formatting, and phrasing to enhance clarity.

docs/diffusion/performance/attention_backends.md

docs/diffusion/performance/cache/cache_dit.md

docs/diffusion/compatibility_matrix.md

docs/diffusion/support_new_models.md

zhaochenyang20 · 2026-02-05T03:26:27Z

@qianyue76 Nice down. Could we leave dLLM part unchange? Since dLLM is in SGLang main endpoint, but diffusion is totally different.

zhaochenyang20 · 2026-02-05T03:28:05Z

docs/advanced_features/server_arguments.md


 ## Diffusion LLM
+
+For detailed documentation on diffusion language models (LLaDA), see [SGLang Diffusion Language Models](../diffusion/language_models/index.md).


Do not change dLLM related docs.

I mean, please put dLLM near LLMs, do not put it under diffusion

zhaochenyang20 · 2026-02-05T03:32:36Z

docs/basic_usage/openai_api_embeddings.ipynb

    "A complete reference for the API is available in the [OpenAI API Reference](https://platform.openai.com/docs/guides/embeddings).\n",
    "\n",
-    "This tutorial covers the embedding APIs for embedding models. For a list of the supported models see the [corresponding overview page](../supported_models/embedding_models.md)\n"
+    "This tutorial covers the embedding APIs for embedding models. For a list of the supported models see the [corresponding overview page](../supported_models/retrieval_ranking/embedding_models.md)"


Nice change to add a middle layer after /supported_models. Please make sure that all the links in our docs that refer to the /supported_models have been clearly changed.

Thanks! I've checked the links and they seem to be correctly pointing to the new subdirectory structure.

zhaochenyang20 · 2026-02-05T03:33:18Z

docs/advanced_features/server_arguments.md


 ## Diffusion LLM
+
+For detailed documentation on diffusion language models (LLaDA), see [SGLang Diffusion Language Models](../diffusion/language_models/index.md).


I mean, please put dLLM near LLMs, do not put it under diffusion

zhaochenyang20 · 2026-02-05T03:36:51Z

docs/diffusion/api/openai_api.md

+  -H "Authorization: Bearer sk-proj-1234567890"
+```
+
+#### Download video content


In the long run, do not use any level 4 title. Just use ###. >=4 level please use ****.

zhaochenyang20 · 2026-02-05T03:38:34Z

docs/diffusion/contributing.md

+
+This guide outlines the requirements for contributing to the SGLang Diffusion module (`sglang.multimodal_gen`).
+
+## 1. Commit Message Convention


If you already use title, then do not add a number in it. Just use:

## Commit Message Convention

python/sglang/multimodal_gen/README.md

…sion/ Closes sgl-project#18089 This commit consolidates all SGLang-Diffusion documentation from: - docs/supported_models/diffusion_models.md - docs/supported_models/diffusion_language_models.md - python/sglang/multimodal_gen/docs/ Into a new unified directory: - docs/diffusion/ Changes: - Create docs/diffusion/ with subdirectories for cache/ and language_models/ - Add new "SGLang Diffusion" section to docs/index.rst - Remove diffusion entries from "Supported Models" section - Add redirect notices to internal docs for backward compatibility - Update python/sglang/multimodal_gen/README.md links to point to new locations

- Create api/ directory for CLI and OpenAI API documentation - Create performance/ directory for optimization docs (caching, profiling, attention) - Merge installation_rocm.md into installation.md - Add performance/index.md overview page - Update all cross-references and internal redirects - Remove duplicate environment variables table from cache_dit.md - Add LLaDA cross-reference in server_arguments.md

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

…ted_models/

Incorporate non-diffusion changes from PR sgl-project#17888: - Reorganize docs/supported_models/ into categorized subdirectories: - extending/ - model extension docs - retrieval_ranking/ - embedding/rerank/classify models - specialized/ - reward models - text_generation/ - generative/multimodal models - Add new model support documentation: - Falcon H1, Hunyuan-Large to generative_models.md - Step3-VL, Qwen3-Omni to multimodal_language_models.md - Reformat support_new_models.md with improved structure - Update index files and references throughout docs - Keep diffusion docs in docs/diffusion/ with short redirect in text_generation/diffusion_language_models.md

- Remove dLLM link from server_arguments.md (keep dLLM near LLMs, not under diffusion) - Move dLLM documentation back to supported_models/text_generation/ - Remove numbered headings in contributing.md (use plain text headings) - Convert #### headings to bold text in diffusion docs (openai_api.md, cli.md, attention_backends.md, cache_dit.md)

zhaochenyang20 · 2026-02-05T18:10:22Z

As I mentioned, we shall merge:

these two: [Docs] Add Falcon H1, Hunyuan-Large, Qwen3-Omni support and update Diffusion usage #17888 [SGLang-Diffusion] Add offline throughput benchmark script for multi-modal models #18154
and finally this: [diffusion] docs: consolidate diffusion documentation into docs #18095

Thanks so much for helping.

zhaochenyang20 · 2026-02-06T22:25:49Z

@qianyue76 since #17888 is merged, and I think we can skip #18154. Could we start to finish this PR right now?

qianyue76 · 2026-02-07T00:51:06Z

@qianyue76 since #17888 is merged, and I think we can skip #18154. Could we start to finish this PR right now?
sure

zhaochenyang20

Basically, moved the docs. LGTM

zhaochenyang20 · 2026-02-10T22:04:52Z

If we have copyed all the contents of docs under multimodal_gen into docs, please just remove them. Not need to left backup.
I am good with the current "hard move", but in the long run, we should have a new subfolder of the docs like:

.. toctree::
   :maxdepth: 1
   :caption: SGLang Diffusion
xxxxxxxxxxxxxxxx

…ion directories Docs have been consolidated into docs/diffusion/. Remove the old multimodal_gen/docs/ and supported_models/image_generation/ directories and update references in basic_usage/diffusion.md accordingly.

zhaochenyang20

Just move the files. Correct to me.

Update links to point to new docs/diffusion/ structure after consolidation. References: - PR #18095: sgl-project/sglang#18095 - Issue #18652: sgl-project/sglang#18652 Changes: - Updated multimodal_gen/docs/* links to docs.sglang.io/diffusion/* - Updated support_matrix.md -> compatibility_matrix.html - Updated install.md -> installation.html - Updated cli.md -> api/cli.html - Updated profiling.md -> performance/profiling.html Co-authored-by: Cursor <cursoragent@cursor.com>

…project#18095) Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com> Co-authored-by: JiaxinD <djx2048@gmail.com>

qianyue76 requested review from mickqian and yhyang201 as code owners February 2, 2026 05:33

github-actions bot added documentation Improvements or additions to documentation amd diffusion SGLang Diffusion labels Feb 2, 2026

gemini-code-assist bot reviewed Feb 2, 2026

View reviewed changes

qianyue76 mentioned this pull request Feb 2, 2026

[Doc] Consolidate Diffusion Documentation into a Unified Directory #18089

Closed

4 tasks

github-actions bot added the Multi-modal multi-modal language model label Feb 4, 2026

zhaochenyang20 requested changes Feb 5, 2026

View reviewed changes

qianyue76 and others added 9 commits February 5, 2026 13:35

Update docs/diffusion/support_new_models.md

80fe026

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

Update docs/diffusion/support_new_models.md

7323682

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

Update docs/diffusion/compatibility_matrix.md

5874de7

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

Update docs/diffusion/performance/cache/cache_dit.md

30fca37

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

[diffusion] docs: add backward compatibility redirects in docs/suppor…

c82d808

…ted_models/

[diffusion] docs: add MUSA/Moore Threads GPU installation guide

bae5657

qianyue76 force-pushed the doc/consolidate-diffusion-docs branch from 0cb7c51 to bae5657 Compare February 5, 2026 05:41

qianyue76 requested a review from zhaochenyang20 February 5, 2026 06:46

zhaochenyang20 mentioned this pull request Feb 5, 2026

[Docs] Add Falcon H1, Hunyuan-Large, Qwen3-Omni support and update Diffusion usage #17888

Merged

5 tasks

zhaochenyang20 approved these changes Feb 7, 2026

View reviewed changes

zhaochenyang20 added the run-ci label Feb 7, 2026

qianyue76 force-pushed the doc/consolidate-diffusion-docs branch from eff453a to bae5657 Compare February 9, 2026 11:54

Merge branch 'main' into doc/consolidate-diffusion-docs

f97629d

github-actions bot added the mthreads label Feb 11, 2026

zhaochenyang20 approved these changes Feb 12, 2026

View reviewed changes

hnyls2002 merged commit f06ab17 into sgl-project:main Feb 12, 2026
58 of 65 checks passed

zhaochenyang20 mentioned this pull request Feb 12, 2026

[Docs] Reaftoring SGLang DIffusion Docs #18652

Open

JiaxinD mentioned this pull request Feb 12, 2026

Fix outdated SGLang diffusion documentation links lm-sys/lm-sys.github.io#314

Merged

3 tasks


		## Diffusion LLM

		For detailed documentation on diffusion language models (LLaDA), see [SGLang Diffusion Language Models](../diffusion/language_models/index.md).


		This guide outlines the requirements for contributing to the SGLang Diffusion module (`sglang.multimodal_gen`).

		## 1. Commit Message Convention

Conversation

qianyue76 commented Feb 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Overview

Before

After

Key Changes

Files Changed

Motivation

Modifications

Accuracy Tests

Benchmarking and Profiling

Checklist

Review Process

Uh oh!

gemini-code-assist bot commented Feb 2, 2026

Summary of Changes

Highlights

Footnotes

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

zhaochenyang20 commented Feb 5, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

zhaochenyang20 commented Feb 5, 2026

Uh oh!

zhaochenyang20 commented Feb 6, 2026

Uh oh!

qianyue76 commented Feb 7, 2026

Uh oh!

zhaochenyang20 left a comment

Choose a reason for hiding this comment

Uh oh!

zhaochenyang20 commented Feb 10, 2026

Uh oh!

zhaochenyang20 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

qianyue76 commented Feb 2, 2026 •

edited

Loading