llama-1b-multimedqa-c100b25-kcenter-raw

A fine-tuned 1B-parameter Llama-family checkpoint adapted for medical or multimodal question-answering tasks, published on Hugging Face by the account camiellia. The naming pattern suggests it was produced via a k-center coreset sampling procedure (c100b25-kcenter-raw) applied during fine-tuning on a multi-domain medical QA dataset. Source: supabase.