stanfordmimi

/

MedVAL-4B-GGUF

like 2

Follow

Stanford MIMI Group 57

This PR improves the model card for MedVAL-4B by enhancing its discoverability and accuracy on the Hugging Face Hub:

• Adds pipeline_tag: text-classification: The model's core function is to assess and validate AI-generated medical text, assigning error categories and risk levels. This fits perfectly with a text classification pipeline, making the model more discoverable for relevant tasks.
• Corrects base_model: The "Model Details" section in the model card clearly states that MedVAL-4B is a "Transformer-based language model (Qwen3-4B)". This indicates that Qwen/Qwen3-4B is the true base model it was fine-tuned from. The previous base_model value was self-referential, and this update provides accurate model lineage information.

No changes were made to the paper links as they already point to the arXiv paper, and per instructions, existing arXiv links do not need to be replaced with Hugging Face Paper page links.