NisalDeZoysa/Qwen-4B-FirstAid-LLM

Uploaded model

• Developed by: Team DataMavericks from Faculty of Engineering, University of Ruhuna
• License: apache-2.0
• Finetuned from model : unsloth/Qwen3-4B-unsloth-bnb-4bit

🩺 FirstAidLLM – Fine-tuned Qwen 4B for First Aid Instructions

📌 Overview

FirstAidLLM is a domain-specialized large language model fine-tuned on the First Aid Instructions Dataset using Qwen 4B as the base model. The goal is to accurately answer medical first aid-related questions in a concise, safe, and reliable way.

🚑 Whether it’s choking, burns, or CPR instructions, FirstAidLLM provides fast and accurate guidance.

✨ Features

✅ Fine-tuned Qwen3-4B using LoRA + 4-bit quantization ✅ Specialized in first aid instructions ✅ Optimized for low GPU memory usage ✅ Compatible with Hugging Face Transformers & TRL SFTTrainer ✅ Easy to deploy on AWS, SageMaker, or local environment

🗂 Dataset

We used the FirstAidInstructionsDataset from Hugging Face.

Field Description question User's first aid-related question answer Accurate instructions to handle the situation category Type of emergency (e.g., choking, CPR) ⚡ Model Architecture Component Details Base Model Qwen3-4B (UnsLoTH) Fine-tuning LoRA (PEFT) Quantization 4-bit (bnb-4bit) Trainer TRL SFTTrainer Framework PyTorch + Transformers