rednote-hilab

/

dots.ocr.base

like 9

Follow

rednote-hilab 639

This PR significantly improves the model card for rednote-hilab/dots.ocr by:

• Updating library_name: Changed the library_name in the metadata from dots_ocr to transformers. This is crucial as the model uses transformers.AutoModelForCausalLM and transformers.AutoProcessor, enabling the "How to use" widget on the Hub for easier adoption.
• Adding prominent links: Introduced new badges at the top for the paper, GitHub repository, and live demo (project page) for better discoverability. The existing live demo link in the text has been replaced by the badge. The X (Twitter) link from the GitHub README has also been added.
• Syncing content with GitHub README:
  • Updated the "News" section with the latest release information.
  • Revised the "Download Model Weights" section to include the ModelScope option.
  • Refreshed the "vLLM inference" instructions under "Deployment" to reflect official vLLM integration (v0.11.0+) and simplified usage.
  • Added a new "Huggingface inference with CPU" section.
  • Updated the "Document Parse" section with the correct --num_thread argument and instructions for Transformers-based parsing.

These changes ensure the model card is up-to-date, more accurate, and more user-friendly, providing clearer guidance for researchers and users.