Google Model Card Toolkit (Python library) for easier Model Transparency Reporting

Follow the full discussion on Reddit.
Google AI recently released a toolkit to generate documentation about their model's performance and intended use case. I personally think the Google Model Cards are a good first step for creating formal standards for evaluating machine learning models beyond just accuracy. By being more rigorous about how we evaluate models, we can instill trust in a model’s real-world performance because we know it has been evaluated thoroughly in many different situations. Many organizations, such as OpenAI, have adopted the Model Card framework or something similar.