NVIDIA-NeMo / NVIDIA-NeMo/Curator

[Docs] Repair audio API examples and add workflow selection guidance

Open
#2,162 0 comments 0 reactions 1 assignee View on GitHub

@lbliii is already working on this.

Since Jul 2, 2026.

documentation needs-follow-up r1.3.0
Dominant language
Python
Stars
1.8k
Forks
327
Avg merge
4d 5h
Merged PRs (30d)
30

Description

Parent tracker: NVIDIA-NeMo/Curator#2118

Goal

Repair stale audio API examples and add task-oriented guidance that helps users choose the correct 26.07 audio workflow.

Problems found

The new audio-tagging and long-form-cutting tutorials are sound, but they link into older supporting pages containing 27 stale import occurrences, principally:

  • nemo_curator.stages.audio.inference.asr_nemo
  • nemo_curator.stages.audio.metrics.get_wer

The current implementations live under inference.asr.asr_nemo and metrics.wer, and some helpers referenced by older pages no longer exist.

The landing pages also list several similar workflows without explaining when to choose basic ASR curation, AudioDataFilterStage, audio tagging, ALM fixed-window construction, or long-form physical cutting and WebDataset packaging.

Timing

Broken imports are release scope and should be corrected before the 26.07 docs publication tag. The workflow-selection presentation can land in the same PR because it is small and directly supports the two new tutorials.

Acceptance criteria

  • Inventory every audio code block and replace removed module paths and symbols with current APIs.
  • Verify examples in getting started, loading, ASR, quality assessment, text integration, and tutorial pages.
  • Add a concise workflow comparison to the audio overview or process-data overview.
  • Distinguish metadata/window construction from physical snippet extraction and packaging.
  • Preserve cross-links to NVIDIA-NeMo/Curator#2135 and #2136 content after their integration conflict is resolved.
  • Run static import checks, fern check, and the generated-reference broken-link job.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.