Instructions to use nagireddy5/medgemma_Q3_K_M_Edge with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nagireddy5/medgemma_Q3_K_M_Edge with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="nagireddy5/medgemma_Q3_K_M_Edge",
	filename="medgemma-Q3_K_M.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use nagireddy5/medgemma_Q3_K_M_Edge with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M
# Run inference directly in the terminal:
llama-cli -hf nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M
# Run inference directly in the terminal:
llama-cli -hf nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M
# Run inference directly in the terminal:
./llama-cli -hf nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M

Use Docker

docker model run hf.co/nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M

LM Studio
Jan
Ollama
How to use nagireddy5/medgemma_Q3_K_M_Edge with Ollama:
```
ollama run hf.co/nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M
```

Unsloth Studio new

How to use nagireddy5/medgemma_Q3_K_M_Edge with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for nagireddy5/medgemma_Q3_K_M_Edge to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for nagireddy5/medgemma_Q3_K_M_Edge to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for nagireddy5/medgemma_Q3_K_M_Edge to start chatting

Docker Model Runner
How to use nagireddy5/medgemma_Q3_K_M_Edge with Docker Model Runner:
```
docker model run hf.co/nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M
```

Lemonade

How to use nagireddy5/medgemma_Q3_K_M_Edge with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull nagireddy5/medgemma_Q3_K_M_Edge:Q3_K_M

Run and chat with the model

lemonade run user.medgemma_Q3_K_M_Edge-Q3_K_M

List all available models

lemonade list

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

Check out the documentation for more information.

🏥 ArogyaNet-AI

Federated Clinical Intelligence for Rural India

Built for the MedGemma Impact Challenge 2026

🌍 Overview

ArogyaNet-AI is an offline-first, federated clinical intelligence platform designed specifically for rural healthcare environments.

India has over 600 million people living in rural areas, where access to specialists, diagnostics, and digital infrastructure is limited. Rural doctors often work without second opinions, rely on paper-based records, and face unstable internet connectivity.

ArogyaNet-AI transforms isolated clinics into intelligent, connected healthcare hubs.

🚨 Problem Statement

1️⃣ No Specialist Backup

Rural doctors often rely solely on personal experience without access to expert consultation.

2️⃣ Diagnostic Gaps

High number of dermatology and general cases with limited specialist access.

3️⃣ Paper-Based Records

Patient histories stored in notebooks → no structured data, no analytics.

4️⃣ Static Care Plans

Generic diet sheets not tailored to patient condition or local food availability.

5️⃣ Unstable Internet

Most AI healthcare systems assume reliable connectivity — rural systems cannot.

💡 Our Solution

ArogyaNet-AI is a multi-agent clinical intelligence ecosystem consisting of:

🧠 Doctor Portal (AI-assisted triage & expert learn)
📱 Offline-First Nurse Mobile App
🔍 Deep Research Agent (Multimodal analysis)
📊 Admin Analytics Dashboard (Disease spike detection)
📞 AI Voice Calling Agent (Automated appointment booking & follow-ups)
🗂 Federated Knowledge Sharing System

🧠 Core Technologies

🔬 MedGemma (Fine-Tuned)

We use MedGemma, a fine-tuned medical version of Google’s Gemma model trained on healthcare datasets for expert-level reasoning.

⚡ Q3_K_M Quantized Edge Model

To support rural connectivity constraints:

Q3 → 3-bit quantization
K_M → Medium k-quant method (llama.cpp framework)
Runs locally on edge devices
No internet required

This allows general medical reasoning to function completely offline.

🧩 Multi-Agent Architecture

MedGemma
LangGraph orchestration
Vector database (knowledge storage)
Multimodal processing (X-ray, audio, PDFs)
Offline-first event caching
Secure anonymized federated sharing

👨‍⚕️ Doctor Features

AI-assisted triage
Real-time second opinions
Voice-to-structured clinical notes
Digital prescriptions & lab ordering
Expert Learn knowledge graph
Strict search within hospital records
Deep Research multimodal analysis

👩‍⚕️ Nurse Features (Offline-First)

Event-based health camp data collection
AI Scan for bulk lab report extraction
Offline vitals logging
Secure local storage
Automatic sync when connectivity returns
Image-based skin analysis with queued processing

👩‍🌾 Patient Features

AI severity assessment & specialist matching
Intelligent appointment scheduling
Document upload (X-ray, reports)
Private AI skin assessments
Personalized diet plans
Regional language summaries
AI voice follow-up reminders

🏥 Admin Features

Real-time dashboard monitoring
Disease spike detection across villages
Camp-level data streaming
Inventory management
Early intervention planning

🌐 Why It Matters

ArogyaNet-AI is designed for:

Remote rural clinics
Health camps
Low-connectivity environments
Resource-constrained hospitals

Healthcare intelligence should not depend on geography.

🚀 Impact Potential

By scaling ArogyaNet-AI across rural India:

Reduce misdiagnosis rates
Enable early disease detection
Improve health literacy
Digitize rural healthcare data
Build a federated rural intelligence network

🏆 Challenge Submission

Submitted to:

The MedGemma Impact Challenge 2026
Build human-centered AI applications using MedGemma and Google’s Health AI Developer Foundations (HAI-DEF).

👥 Team

Nagi Reddy – AI Architect & Systems Integration
Harshith – Video Production & Presentation
Venkatesh – Research & Resource Support

📌 Vision

From isolated rural rooms to intelligent, connected healthcare ecosystems.

ArogyaNet-AI — Care Never Stops.

license: mit

Downloads last month: 1

GGUF

Model size

4B params

Architecture

gemma3

Hardware compatibility

3-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support