This model is a fine-tuned version of LLaMA-3 8B Instruct using QLoRA (4-bit quantization + LoRA adapters) on the MedQuad Medical QnA Dataset.
It is designed to answer medical domain questions across various categories like treatment, symptoms, causes, prevention, inheritance, etc.

Model Details

Model Description

Base model: meta-llama/Meta-Llama-3-8B-Instruct
Fine-tuning method: QLoRA (4-bit quantization with LoRA adapters)
Task: Medical Question Answering (Instruction-tuned style)
Languages: English
Framework: 🤗 Transformers, PEFT, TRL
Quantization: 4-bit (nf4, bfloat16 compute)
License: Llama 3 license

Developers

Developed by: Arush Pettem
Dataset: keivalya/MedQuad-MedicalQnADataset

Model Sources

Repository: [Your Hugging Face repo link]
Paper: "MedQuAD: Medical Question Answering Dataset"
Demo: (Optional if you make a Gradio Space)

Uses

Direct Use

Answering medical questions in categories such as treatment, symptoms, causes, prevention, outlook, etc.
Educational and research purposes in healthcare QA systems.

Downstream Use

Integration into healthcare chatbots.
Fine-tuning on domain-specific sub-corpora (e.g., cardiology QnA).
Evaluation for explainable AI in medical NLP.

Out-of-Scope Use

⚠️ This model is not a substitute for professional medical advice. It should not be used for clinical decision-making or diagnosis.

Bias, Risks, and Limitations

Bias: Model inherits potential biases from MedQuad and the LLaMA base model.
Risks: Incorrect or incomplete medical answers may mislead users if used in real-world clinical contexts.
Limitations: Trained on static QA pairs, so may not generalize to open-ended patient conversations.

Recommendations

Use in controlled, educational, or research settings only.
Always validate outputs with trusted medical sources.

How to Get Started with the Model

'''https://huggingface.co/Arushp1/llama3-medquad-qlora'''

python

from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
model = AutoModelForCausalLM.from_pretrained("Arushp1/llama3-medquad-qlora")
tokenizer = AutoTokenizer.from_pretrained("Arushp1/llama3-medquad-qlora")

pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

query = "What are the symptoms of asthma?"
print(pipe(query, max_new_tokens=100))