Content Guidelines 🤗

💡 Also read the Hugging Face Code of Conduct which gives a general overview and states our standards and how we wish the community will behave.

🗓 Effective Date: August 10, 2022

Hugging Face’s purpose is to help the community work together to advance Open, Collaborative, and Responsible Machine Learning. Hugging Face’s achievements are only possible thanks to our awesome community.

We value these relationships and aim to maintain an environment where people feel welcome and supported and get the most out of their contributions and experiences. In order to do that, we must develop our understanding of the downstream consequences of our choices and foster foresight to make responsible decisions under the auspices of our platform.

This document aims to support Hugging Face’s actions to protect our community on our platform, allowing them to flag types of content with a higher risk of harming people and requiring additional attention.

🤔 What's the content?

The content on the Hugging Face platform is made up of:

🤖 Technical content:

  • Datasets
  • Models
  • Spaces/Demos

🧑 Human content:

  • Discussions & comments under the community tabs

🛎️ How can you report content?

The community tab allows you to bring issues to the attention of the Hugging Face community by opening discussions within each project, proposing Pull Requests to modify its content, and suggesting ways to address the problems. If you encounter harmful content, you can directly flag the repository via the "report" button on the hub. This action will open a public discussion in the community tab and ping the Hugging Face team, who will act accordingly to the following rules.

🕵️ How do we examine and measure inappropriate content?

We have a case-by-case measurement approach based on a triage system that allows us to classify the content and respond accordingly.

Timeline:

Once the report has been filed, best efforts from the Hugging Face team will be made to ensure your request is treated as soon as possible.

Technical content:

The following three measures are not necessarily sequential and can occur independently. Depending on the severity of the content, we may find ourselves in the situation of acting quickly.

🧐 Hey… friendly warning 🤨 Ouch… reaching the limit 😡 Ok… enough is enough!
We open a public discussion on the community tab raising the issue and asking for feedback. We send a written warning to the author clarifying the flagged content and consequences of failing to address the issue. We open a public discussion on the community tab raising the issue and asking for feedback. We send a written warning to the author clarifying the flagged content and consequences of failing to address the issue. Additionally, we evaluate the possibility of gating access to the content and downgrading the content's visibility on the platform. We open a public discussion on the community tab raising the issue and asking for feedback. We send a written warning to the author clarifying the flagged content and consequences of failing to address the issue. Additionally, we evaluate the possibility of blocking access to the content and downgrading the content's visibility on the platform.

Human content:

We do not tolerate the following human content on our platform:

  • Content promoting discrimination (see our Code of Conduct) & Hate speech.
  • Harassing, demeaning, or bullying
  • Sexual content involving minors, or used or created for harassment, bullying, or without explicit consent of the people represented.
  • Spam: a) it promotes a product or service, or b) excessive bulk activity
  • Other: a problem not listed above that requires our attention

Out of consideration for other users of the Hub, we also request that the following kinds of content be clearly marked (for example, in the repository's data or model card), and when possible to have the Hub users opt in to see the content (for example, by using the Hub's model or dataset repository gating function).

  • Un-requested violent content
  • Un-requested sexual content

Finally, we hold consent as a fundamental value at Hugging Face. We ask that you do not use our services to share or upload other people's data, such as their images, media recordings, or text, without their explicit consent. As we as a community develop better norms around trainng data and AI-related rights, we also ask that you take reasonable steps to honor other people's wishes regarding systems trained primarily on their data.

Consequences:

🧐 Hey… friendly warning 🤨 Ouch… reaching the limit 😡 Ok… enough is enough!
A private, written warning from the Hugging Face team, providing clarity around the nature of the content and an explanation of why the behavior was inappropriate. A warning with consequences for continued behavior. Violating these terms may lead to a temporary ban from any sort of public interaction within the community tab on the platform. A permanent ban from any sort of public interaction within the community tab on the platform.

👷‍♀️ Intellectual property rights infringement:

Digital Millennium Copyright Act (DMCA) takedown is a tool for copyright holders to get user-uploaded material that infringes their copyrights taken down off websites.

At Hugging Face, we comply with the DMCA! If you have any claims that any content on our Website violates or infringes your intellectual property rights, you may send your complaint to dmca@huggingface.co with detailed and accurate information supporting your claim.

You also represent and warrant that you will not knowingly provide misleading information to support your claim.

🛠️ Additional tools:

  • Authors of a Discussion or a Pull Request can edit the discussion's title.
  • Repository owners can choose to hide a community comment.