Skip to content

docs: add FastVideo FastH3 fast generation guide for MiniMax H3 - #1690

Open
lin-bot23 wants to merge 2 commits into
Comfy-Org:mainfrom
lin-bot23:docs/fastvideo-fasth3
Open

lin-bot23 wants to merge 2 commits into
Comfy-Org:mainfrom
lin-bot23:docs/fastvideo-fasth3

Conversation

@lin-bot23

Copy link
Copy Markdown
Contributor

workflow_templates v0.11.61/v0.11.62 (Comfy-Org/workflow_templates#1264, #1265) added two FastVideo FastH3 templates for MiniMax H3: video_fastvideo_fasth3_t2v and video_fastvideo_fasth3_i2v (openSource, minComfyUIVersion 0.36.0).

This adds a new tutorial page tutorials/video/minimax/minimax-h3-fastvideo.mdx covering:

  • What FastH3 is: a DMD2-distilled MiniMax H3 checkpoint generating synchronized video and audio in 8 sampling steps, reusing the H3 text encoder and VAEs
  • Scope note: t2va and fl2va only, Ref2VA was not distilled
  • Both workflows with previews, download cards, prompting tips, and input material
  • Model downloads (FastVideo diffusion model + shared H3 text encoder/VAEs) and storage tree
  • Workflow structure notes (BlockSparseAttention VSA settings, MiniMaxH3SigmaShift, 17k+5 frame grid)

Nav entries added to the MiniMax H3 group in all 4 languages. Comfy Cloud cards are commented out until the templates are available in Comfy Cloud.

@coderabbitai

coderabbitai Bot commented Sep 16, 2026

Copy link
Copy Markdown

Review Change StackReview Change Stack

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Advanced

Run ID: bd83849a-dd7e-4ac5-86e6-18a808ee93b3

📥 Commits

Reviewing files that changed from the base of the PR and between c469616 and 434f993.

📒 Files selected for processing (4)
  • ja/tutorials/video/minimax/minimax-h3-fastvideo.mdx
  • ko/tutorials/video/minimax/minimax-h3-fastvideo.mdx
  • tutorials/video/minimax/minimax-h3-fastvideo.mdx
  • zh/tutorials/video/minimax/minimax-h3-fastvideo.mdx

Included review availability: Your plan provides up to 2 included reviews per hour; 1 remains after this review.


📝 Walkthrough

Walkthrough

FastH3 documentation was added in English, Japanese, Chinese, and Korean. The pages describe text-to-video and first/last-frame image-to-video workflows, model setup, prompt guidance, and workflow settings. Navigation now links to the new pages.

Changes

MiniMax H3 FastVideo Documentation

Layer / File(s) Summary
English FastH3 tutorial
tutorials/video/minimax/minimax-h3-fastvideo.mdx
Added FastH3 text-to-video and first/last-frame image-to-video workflows, model downloads, prompt guidance, and workflow settings.
Localized FastH3 tutorials
ja/tutorials/video/minimax/minimax-h3-fastvideo.mdx, zh/tutorials/video/minimax/minimax-h3-fastvideo.mdx, ko/tutorials/video/minimax/minimax-h3-fastvideo.mdx
Added Japanese, Chinese, and Korean translations of the FastH3 tutorial.
Multilingual tutorial navigation
docs.json
Added the tutorial to the English, Chinese, Japanese, and Korean MiniMax H3 navigation lists.

Priority: ⬇️ Low

Merge Risk: ⚪ Minimal · up to 434f9

The PR adds the FastH3 documentation and multilingual navigation without an identified merge-blocking risk.

🚥 Pre-merge checks | ✅ 2
✅ Passed checks (2 passed)
Check name Status Explanation
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.
✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests
✨ Simplify code
  • Create PR with simplified code

Comment @coderabbitai help to get the list of available commands.

Align with the template gallery naming ("FastVideo FastH3: Text to Video" /
"Image to Video") and the LTX-2.5 page convention of leading with the model
name. sidebarTitle simplifies to FastVideo FastH3.
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.

1 participant