Skip to content

RFC Q1: training-data audit for oh-sheet-ml-pipeline weights #114

Description

@shoyu-ramen

Tracks Q1 of the ml-pipeline publishing RFC. Spun out of #107.

What's needed

A complete catalogue of the datasets used to train the oh-sheet-ml-pipeline engraver model, with the redistribution license for each. Specifically:

  • Every dataset (or scraped corpus) the model was trained or fine-tuned on.
  • The license / terms-of-use for each dataset, with a primary-source link.
  • Whether those terms permit redistribution of model weights derived from the data (commercial / non-commercial / research-only / unclear).
  • If any dataset blocks redistribution: whether re-training or fine-tuning on a clean corpus is feasible, and what that would cost.

Why this gates the publishing decision

If the audit surfaces a dataset whose license forbids redistribution of derived weights, then options A (open-source repo + image) and B (image-only) from #107 are blocked until the model is retrained on a clean corpus — regardless of how Q3 (security review) and Q4 (maintenance commitment) shake out.

The plugin protocol C (docs/engraver-protocol/) is not blocked by this question — it's a documented HTTP contract, no weights involved.

How to help

Anyone with experience auditing training-data provenance for music-AI models — or with knowledge of which public music datasets have clean redistribution terms — please comment.

The actual audit work has to happen in the private oh-sheet-ml-pipeline repo, so progress will be summarized here rather than discussed line-by-line.

Refs: #107

Metadata

Metadata

Assignees

No one assigned

    Labels

    documentationImprovements or additions to documentationhelp wantedExtra attention is neededquestionFurther information is requested

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions