Lightweight medical assistant chatbot fine-tuned with QLoRA on Mistral-7B-Instruct using the ruslanmv/ai-medical-chatbot dataset. Includes a Streamlit chat UI and a Colab-ready fine-tuning notebook.

⚠️ Disclaimer: Educational triage guidance only; not medical advice or diagnosis.

1. Open the notebook notebooks/01_finetune_mistral_medical_QLoRA.ipynb in Colab/Kaggle.
2. Run all cells → saves LoRA adapters to artifacts/lora/medical_chatbot_lora/.
3. Install deps and run Streamlit app:

   pip install -r requirements.txt
   streamlit run app.py

Update config/config.yaml to change model or generation settings.