aparnavirtuonai/mistral-medqa: Medical QA Fine-Tuned Model

This model is a 7 billion parameter language model, fine-tuned from the mistralai/Mistral-7B-v0.1 base model. Its primary purpose is to excel in medical question-answering by leveraging domain-specific knowledge.

Key Capabilities

• Medical Question Answering: Specifically trained to understand and respond to medical queries.
• Healthcare Chatbot Systems: Suitable for integration into conversational AI for healthcare contexts.
• Educational and Research: Can be used for learning and investigative purposes within the medical field.
• Fine-Tuned Performance: Achieves 0.70 accuracy on BoolQ (General QA) and 0.69 accuracy on PubMedQA (Medical QA), indicating stable general performance alongside domain adaptation.

Training Details

The model underwent Supervised Fine-Tuning (SFT) using LoRA (Low-Rank Adaptation) with PEFT. It was trained for 3 epochs on the MedQA dataset, utilizing a learning rate of 2e-5 and a maximum sequence length of 512 tokens.

Good For

• Developing applications that require accurate responses to medical questions.
• Building specialized chatbots for healthcare information.
• Supporting medical education and research initiatives.

Note: This model is not intended for real-world medical diagnosis without professional supervision.