Upload FireRedASR v2 models (AED and CTC)

[csukuangfj]

Closes #3219

cc @dfrrk

Download link: sherpa-onnx-fire-red-asr2-zh_en-int8-2026-02-26.tar.bz2

You can try it at
https://huggingface.co/spaces/k2-fsa/automatic-speech-recognition

---

Its usage is identical with
https://k2-fsa.github.io/sherpa/onnx/FireRedAsr/pretrained.html#sherpa-onnx-fire-red-asr-large-zh-en-2025-02-16-chinese-english

You don't need to change any code.

Summary by CodeRabbit

• Chores
  • Updated build automation and deployment workflows with improved tooling for model packaging and publishing.

[gemini-code-assist]

Summary of Changes

Hello @csukuangfj, 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, while titled to introduce new FireRedASR v2 models, primarily contains a minor update to a GitHub Actions test script. The change involves migrating a model download command from wget to curl -SL -O within the Node.js addon npm test workflow. This modification improves the robustness of the download process for an existing medasr-ctc model, rather than directly adding the new FireRedASR v2 models to the codebase.

Highlights

• Intended Model Upload: This pull request is titled to upload new FireRedASR v2 models (AED and CTC) for both Chinese and English languages, as indicated by the pull request description and associated download link.
• Test Script Update: The actual code change in this pull request updates a GitHub Actions test script, replacing the wget command with curl -SL -O for downloading the sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2 model archive.

Changelog

• .github/scripts/test-nodejs-addon-npm.sh
  • Replaced wget with curl -SL -O for downloading the sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2 model archive.

Ignored Files

• Ignored by pattern: .github/workflows/** (1)
  • .github/workflows/upload-models.yaml

Activity

• The pull request was created by csukuangfj, linking to issue [Help wanted] 请加入对小红书出的2代模型FireRedASR2S 这个模型的支持，感谢！ #3219 and mentioning @dfrrk.
• A download link for the new FireRedASR v2 models was provided in the description, along with a Hugging Face space for testing and a screenshot.

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.

Footnotes

1. 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 changes a download command in a test script from wget to curl. However, there is a major discrepancy between this change and the PR's title and description, which refer to 'Upload FireRedASR v2 models'. This is misleading and should be corrected to reflect the actual change. I have also provided a suggestion to make the new curl command more robust.

[gemini-code-assist]

Using -O relies on the URL's path to determine the output filename. While it works here, it's more robust to explicitly specify the output file with -o <filename>. This decouples the script from the URL structure and ensures the filename matches what subsequent commands like tar and rm expect.

Suggested change

	curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2
	curl -SL -o sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2 https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

Configuration used: defaults

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between cb0828a and 8d1cadd.

📒 Files selected for processing (2)

• .github/scripts/test-nodejs-addon-npm.sh
• .github/workflows/upload-models.yaml

---

📝 Walkthrough

Walkthrough

This PR introduces support for FireRedASR2 ASR models through new disabled GitHub Actions packaging and publishing jobs for int8 and fp32 variants, updates publishing lists with four new model identifiers, adjusts Git LFS tracking to include "*.weights" files, and replaces a wget command with curl in the test script.

Changes

Cohort / File(s)	Summary
Test Script Update .github/scripts/test-nodejs-addon-npm.sh	Replaces wget download with curl for fetching tarball in the non-streaming ASR Google MedASR CTC section.
Workflow Model Jobs .github/workflows/upload-models.yaml	Adds two new gated FireRedASR2 CTC model packaging jobs (int8 and fp32 variants), each downloading ONNX assets and archiving to tar.bz2. Jobs are disabled via if: false.
Workflow Publishing Steps .github/workflows/upload-models.yaml	Adds four new FireRedASR2 model identifiers to Publish to huggingface and Publish to modelscope sequences; adjusts Git LFS tracking to include "*.weights" files; removes a prior upload-artifact step.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• #2453: Adds gated model packaging steps for new ASR models in the same workflow
• #3038: Extends upload-models workflow with FireRed ASR model jobs and publishing configurations
• #2470: Modifies test-nodejs-addon-npm.sh to adjust model download and extraction steps

Suggested labels

size:M

Poem

🐰 A rabbit hops with joy and cheer,
FireRedASR2 models now are here!
From wget swapped to curl so spry,
New packaging jobs reach for the sky,
With int8 and fp32 aligned in song! ✨

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

Tip

Try Coding Plans. Let us write the prompt for your AI agent so you can ship faster (with fewer bugs).
Share your feedback on Discord.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}