YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

license: apache-2.0 language:

  • en tags:
  • medical
  • gemma
  • lora
  • unsloth
  • healthcare pipeline_tag: text-generation

MedQA Gemma 1B LoRA

A medical question-answering LoRA fine-tuning project using Gemma 3 1B and Unsloth.

MedQA Gemma 1B LoRA

A medical question-answering LoRA fine-tuning project using Gemma 3 1B and Unsloth.

Model Overview

This model was fine-tuned on a subset of the MedQA-USMLE dataset using:

  • Unsloth
  • LoRA
  • Google Colab
  • Tesla T4 GPU

The goal of this project is to explore medical instruction tuning and lightweight domain adaptation for healthcare-related question answering.

Base Model

unsloth/gemma-3-1b-it

Dataset

GBaker/MedQA-USMLE-4-options

Training Details

  • Training Method: LoRA
  • GPU: Tesla T4
  • Framework: Unsloth
  • Max Steps: 30
  • Training Samples: 500
  • Precision: float32

Example Prompt

Question: A pregnant woman has dysuria without fever or flank pain.

Options: A. Doxycycline B. Nitrofurantoin C. Vancomycin D. Oseltamivir

Example Output

B. Nitrofurantoin

Intended Use

This project is intended for:

  • AI experimentation
  • educational purposes
  • medical LLM fine-tuning research

Limitations

This model is NOT intended for:

  • real clinical diagnosis
  • medical treatment decisions
  • production healthcare deployment

It was trained on a limited subset for experimentation purposes only.

Tools Used

  • Hugging Face Transformers
  • Unsloth
  • TRL
  • PEFT
  • LoRA
  • Google Colab

Author

Ayman Mohammed

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support