README.md

metadata

title: README
emoji: 🏭
colorFrom: gray
colorTo: red
sdk: static
pinned: false
thumbnail: >-
  https://cdn-uploads.huggingface.co/production/uploads/6ae0fe779baaa10b775edbfd/mlI7tZfPDrZml8F4rSN0l.png

LexData Labs

LexData Labs builds training data for computer vision systems that run in the physical world: energy infrastructure, manufacturing lines, vehicle and property inspection, and oil and gas operations.

This org hosts our open datasets, benchmarks, and tools.

Lexiqal

LexData Industrial Quality Assessment Leaderboard (coming soon)

Most vision benchmarks test models on everyday photos. Lexiqal tests them on the conditions that break industrial deployments: glare, night and thermal imagery, weather, occlusion, motion blur, and defects that look a lot like normal variation.

• Open data. Built from public Hugging Face datasets in industrial and mechanical domains. Each source dataset keeps its original license and attribution.
• Held-out test split. A private test split backs the leaderboard so scores stay meaningful.
• Per-condition results. Scores are broken down by domain and condition.

Submission instructions and the first domain release will be posted here.

Links

• Website: lexdatalabs.com
• GitHub: github.com/LexData-Labs

Have a dataset you think belongs in Lexiqal, or found a labeling error? Open a discussion on the relevant repo.