Aikyam-Lab/CURE-MED-7B
Aikyam-Lab/CURE-MED-7B is a 7.6 billion parameter multilingual medical reasoning large language model developed by Aikyam Lab and collaborators. Fine-tuned from Qwen2.5-7B, it utilizes a curriculum-informed reinforcement learning framework to enhance logical correctness and language stability in healthcare applications. This model excels at open-ended medical queries across 13 languages, including underrepresented ones like Amharic, Yoruba, and Swahili. It is specifically designed for robust performance in multilingual medical reasoning tasks.
Loading preview...
CURE-MED-7B: Multilingual Medical Reasoning LLM
CURE-MED-7B is a 7.6 billion parameter large language model developed by Aikyam Lab and collaborators, specialized in multilingual medical reasoning. Built upon the Qwen2.5-7B base model, it addresses the complexities of medical queries across diverse languages.
Key Capabilities
- Multilingual Medical Reasoning: Optimized for open-ended medical questions across 13 languages, including Amharic, Yoruba, Swahili, Bengali, French, Hausa, Hindi, Japanese, Korean, Spanish, Thai, Turkish, and Vietnamese.
- Enhanced Logical Correctness: Utilizes a curriculum-informed reinforcement learning framework, integrating code-switching-aware supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO).
- Robust Performance: Trained and evaluated using CUREMED-BENCH, a high-quality multilingual open-ended medical reasoning benchmark with single verifiable answers.
Good For
- Applications requiring accurate medical reasoning in multiple languages.
- Healthcare systems operating in linguistically diverse environments.
- Research and development in multilingual AI for medicine.
For more details, refer to the project repository and the research paper.