README.md

metadata

title: ModelBrew Dataset Cleaner
emoji: 🧹
colorFrom: yellow
colorTo: red
sdk: static
app_file: index.html
license: apache-2.0
short_description: 90+ quality checks for fine-tuning datasets

ModelBrew Dataset Cleaner

Upload a fine-tuning dataset (.jsonl, .json, .csv) and get a scored issue report — PII with checksum validation, exact/near duplicates, prompt-injection and jailbreak patterns, label errors, truncated responses, and more — plus a cleaned export with critical rows dropped.

Run it locally on datasets of any size: pip install modelbrew-cleaner.

Built by ModelBrew.