The State of Documentation Practices of Third-party Machine Learning Models and Datasets

Abstract

Model stores offer third-party ML models and datasets for easy project integration, minimizing coding efforts. One might hope to find detailed specifications of these models and datasets in the documentation, leveraging documentation standards such as model and dataset cards. In this study, we use statistical analysis and hybrid card sorting to assess the state of the practice of documenting model cards and dataset cards in one of the largest model stores in use today--Hugging Face (HF). Our findings show that only 21,902 models (39.62\%) and 1,925 datasets (28.48\%) have documentation. Furthermore, we observe inconsistency in ethics and transparency-related documentation for ML models and datasets.

arXiv comment7 pages, 4 figures, IEEESoftware format

References in corpus (4)

• Model Cards for Model Reporting
• Challenges in Deploying Machine Learning: a Survey of Case Studies
• Aspirations and Practice of Model Documentation: Moving the Needle with Nudging and Traceability
• Reusable Templates and Guides For Documenting Datasets and Models for Natural Language Processing and Generation: A Case Study of the HuggingFace and GEM Data and Model Cards

Cited by in corpus (1)

• Building an Open AIBOM Standard in the Wild