Robot2050

/

MoM

like 0

This PR significantly enhances the model card for the MoM: Mixtures of Scenario-Aware Document Memories for Retrieval-Augmented Generation Systems framework.

Key improvements include:

• Metadata: Added library_name: transformers and pipeline_tag: text-generation. The transformers library compatibility is evidenced by the config.json (Qwen2ForCausalLM architecture) and transformers_version. This will improve discoverability and enable the automated "How to use" widget on the Hugging Face Hub.
• Comprehensive Overview: Incorporated the paper abstract and "Core Contributions" section directly from the paper and GitHub README, providing a detailed understanding of the framework.
• Target Audience: Included the "Who Should Pay Attention to Our Work?" section from the GitHub README to clarify the relevance for different research and development fields.
• Enhanced Badges: Updated the header with a more complete set of badges from the official GitHub repository, including links to the arXiv paper, Hugging Face paper, and related Hugging Face models.
• Related Models: Added a "Checkpoints" section linking to other models within the MoM family available on the Hugging Face Hub.
• Usage Guidance: Provided a "Quick Start" section directing users to the comprehensive instructions on the official GitHub repository, aligning with the instruction not to create custom code snippets for usage.
• Citation: Added a BibTeX entry for easy citation.

These updates aim to provide a more informative, discoverable, and user-friendly experience for anyone interacting with the MoM framework and its models on the Hugging Face Hub.