Improve model card for Watermark Anything with detailed usage

#1
by nielsr HF Staff - opened

This PR significantly enhances the model card for the facebook/watermark-anything model by:

  • Adding pipeline_tag: image-to-image and relevant tags (watermarking, image watermarking, splicing) to improve discoverability.
  • Including the full abstract of the paper "Watermark Anything with Localized Messages" for a comprehensive overview.
  • Integrating the main illustrative image from the project to visually represent the model's function.
  • Adding the "News" section from the GitHub repository to keep the community updated on model developments.
  • Providing detailed installation instructions and extensive Python code snippets for both single and multiple watermark inference, directly from the original GitHub repository, emphasizing how to load and use the model weights downloaded from the Hub.
  • Clarifying the licensing information for the different model checkpoints as found in the GitHub README.
  • Ensuring all relevant links (Paper, GitHub, Colab, Hugging Face Space) are present and easily accessible.

This update aims to make the model more understandable and easier to adopt for researchers and developers.

pierrefdz changed pull request status to merged

Sign up or log in to comment