kandinskylab

/

Kandinsky-5.0-T2V-Pro-sft-5s

like 3

Follow

Kandinsky Lab 107

This PR significantly improves the model card for the Kandinsky 5.0 family of models by adding crucial metadata and structured content.

Key changes include:

• Metadata: Added pipeline_tag: text-to-video for better discoverability on the Hub. Added library_name: diffusers as the model is officially compatible with the diffusers library, enabling the automated "how to use" widget. The existing license: mit is retained.
• Model Overview: Included a concise introduction to the Kandinsky 5.0 family, highlighting its capabilities in image and video generation.
• Links: Provided direct links to the official research paper, project page, and the GitHub repository.
• Sample Usage: Incorporated a Python code snippet for text-to-video inference, copied directly from the "Quickstart" section of the GitHub README, demonstrating how to use the kandinsky library, which integrates with diffusers.
• Citation: Added the BibTeX entry for the Kandinsky 5.0 paper for proper academic attribution.
• Conciseness: The content has been curated to be focused and relevant for the Hugging Face Hub, directing users to the comprehensive GitHub repository for extensive details, model variations, and further examples.

This update makes the model card more informative, discoverable, and user-friendly for the community.