eunwoneunwon

/

EPISODE-extraction_llama3

like 0

This PR significantly enhances the model card for the SHARE model by providing comprehensive details derived from its associated paper, "SHARE: Shared Memory-Aware Open-Domain Long-Term Dialogue Dataset Constructed from Movie Script".

Key updates include:

• A detailed model description, clarifying its role as a LoRA adapter for meta-llama/Meta-Llama-3.1-8B-Instruct within the EPISODE framework.
• Linking to the official Hugging Face paper page: SHARE: Shared Memory-Aware Open-Domain Long-Term Dialogue Dataset Constructed from Movie Script.
• Adding the appropriate pipeline_tag: feature-extraction to improve discoverability, reflecting its function in managing and extracting shared memories.
• Specifying the base_model, language, datasets, and license in the metadata.
• Including relevant tags such as dialogue, long-term-dialogue, memory, conversational, llama, llm-adapter, and peft.
• Providing a clear usage example for loading and interacting with the LoRA adapter.
• Populating the Bias, Risks, and Limitations section based on the nature of the training data (movie scripts).
• Adding a BibTeX citation for the paper.

Please review and merge this PR to make the model more informative and usable for the community.