CYX1998/

Meissa-SFT

This dataset contains the open-source portion of the SFT (Supervised Fine-Tuning) training data for Meissa, a lightweight 4B-parameter medical multi-modal LLM with full agentic capability.

Total samples: 25,018 Format: ShareGPT (compatible with LLaMA-Factory) Language: English Domain: Medical (radiology, pathology, clinical reasoning)

	Data Sources




	

Framework Dataset Samples Description… See the full description on the dataset page: https://huggingface.co/datasets/CYX1998/Meissa-SFT.