younesgu/medical_chatbot

Llama-2-7B Medical Conversations (QLoRA Finetuned)

This model is a specialized fine-tuned version of Llama-2-7B-chat, optimized for medical-style conversations using 275k real patient–doctor dialogues from Kaggle.
It improves the empathy, clarity, and structure of medical communication but must not be used for real medical diagnosis or treatment.

---

Model Details

Model Description

This model imitates the structure and reasoning style of doctors in conversation with patients.
It was trained using QLoRA to efficiently adapt Llama-2-7B using low-rank adapters and 4-bit quantization.
It is suitable for educational, research, and prototype medical chatbot use cases.

• Developed by: Younes
• Model type: Instruction-tuned conversational medical assistant
• Language(s): English
• Finetuned from model: Llama-2-7B-Chat-HF

---

Uses

Direct Use

The model can be used for:

• Medical-style conversational answering
• Patient triage–style questioning
• Symptom-based reasoning
• Health advice summarization
• FAQ generation in healthcare contexts
• Educational/academic simulations of doctor–patient dialogue

⚠️ The model must NOT be used for real diagnosis or emergency medical advice.

Downstream Use

• Further fine-tuning for medical specialties
• Prototype medical chatbots (non-clinical)
• Synthetic medical conversation generation
• Medical education tools
• Research on health-oriented NLP

Out-of-Scope Use

• Real clinical decision-making
• Prescriptions or medication guidance
• Emergency response
• Deployment in hospitals or regulated health environments
• Automated triage systems without expert validation

---

Recommendations

Users should:

• Treat outputs as non-medical suggestions only.
• Expect occasional hallucinations due to small dataset size.
• Review all generated content manually before use.
• Avoid deploying in any clinical or healthcare-critical project.

---

Framework versions

• PEFT 0.7.1