dataset-audit

Safe, auditable YOLO missing-label recovery | 安全可审计的 YOLO 漏标恢复工具

Quality-aware, read-only MCP server for discovering and selecting African speech corpora, Wolof first, with audited metrics, provenance, licensing, and training-set planning.

Local-first computer vision dataset audit dashboard with a Python engine for leakage, duplicates, blur, imbalance, and evidence-based recommendations.

Audit clinical NLP datasets before training. Offline, deterministic, no cloud model.

Petroleum-engineer-friendly coverage audit of the Equinor Volve dataset from a frozen filesystem catalog.

Companion repository for annotations, back translations, benchmark analyses, and supporting materials from our audit of RWTH-PHOENIX-2014T as a benchmark for German Sign Language translation.

Reproducible audit of the MAD dataset (MAST failure taxonomy): three undocumented taxonomy versions, renumbered codes, and what that means for reported inter-annotator agreement.

Local-first QA and leakage/drift auditor for medical image segmentation datasets.

Checks datasets for privacy leaks, duplicates, split leakage, and missing provenance.