Instructions to use emylton/arogya-ai-full with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use emylton/arogya-ai-full with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="emylton/arogya-ai-full")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("emylton/arogya-ai-full")
model = AutoModelForCausalLM.from_pretrained("emylton/arogya-ai-full")

llama-cpp-python

How to use emylton/arogya-ai-full with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="emylton/arogya-ai-full",
	filename="arogya-ai-fp16.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Inference
Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use emylton/arogya-ai-full with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf emylton/arogya-ai-full
# Run inference directly in the terminal:
llama-cli -hf emylton/arogya-ai-full

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf emylton/arogya-ai-full
# Run inference directly in the terminal:
llama-cli -hf emylton/arogya-ai-full

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 emylton/arogya-ai-full
# Run inference directly in the terminal:
./llama-cli -hf emylton/arogya-ai-full

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 emylton/arogya-ai-full
# Run inference directly in the terminal:
./build/bin/llama-cli -hf emylton/arogya-ai-full

Use Docker

docker model run hf.co/emylton/arogya-ai-full

LM Studio
Jan

vLLM

How to use emylton/arogya-ai-full with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "emylton/arogya-ai-full"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "emylton/arogya-ai-full",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/emylton/arogya-ai-full

SGLang

How to use emylton/arogya-ai-full with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "emylton/arogya-ai-full" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "emylton/arogya-ai-full",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "emylton/arogya-ai-full" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "emylton/arogya-ai-full",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use emylton/arogya-ai-full with Ollama:
```
ollama run hf.co/emylton/arogya-ai-full
```

Unsloth Studio

How to use emylton/arogya-ai-full 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 emylton/arogya-ai-full 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 emylton/arogya-ai-full to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for emylton/arogya-ai-full to start chatting

Docker Model Runner
How to use emylton/arogya-ai-full with Docker Model Runner:
```
docker model run hf.co/emylton/arogya-ai-full
```

Lemonade

How to use emylton/arogya-ai-full with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull emylton/arogya-ai-full

Run and chat with the model

lemonade run user.arogya-ai-full-{{QUANT_TAG}}

List all available models

lemonade list

Arogya AI - Full Model

Arogya (Sanskrit: "perfect health") adalah model AI kesehatan yang dilatih khusus untuk analisis dan prediksi data kesehatan di Kabupaten Maluku Tenggara, Indonesia.

🔗 GitHub Repository: https://github.com/Rafael2022-prog/arogya-ai

📄 Research Paper:

🎯 LoRA Adapter Version: emylton/arogya-health-model (27 MB)

Model Description

Ini adalah full merged model (Llama 3 8B + LoRA adapter) yang siap pakai tanpa perlu download base model terpisah.

Key Features

Ready to use: Tidak perlu base model Llama 3
Ollama compatible: Bisa langsung import ke Ollama
Specialized: Dilatih dengan 10,000+ data kesehatan real dari Maluku Tenggara
Multi-disease: Mendukung 7 penyakit utama (DBD, ISPA, Malaria, Diare, TB, Stunting, Pneumonia)
Geographic coverage: 9 kecamatan di Kabupaten Maluku Tenggara

Training Data

Model dilatih menggunakan 10,000+ records dari 4 sumber data:

LAMPIRAN PROFIL MALUKU TENGGARA 2023 (Excel)
LAMPIRAN PROFIL KESEHATAN MALRA 2024 (Excel)
RENJA 2026 DINKES MALRA (PDF - 24 pages, 14 tables)
RENSTRA DINAS KESEHATAN 2025-2029 (PDF - 97 pages, 74 tables)

Data mencakup 169 indikator kesehatan dari tahun 2020-2029.

Model Details

Base Model: Meta-Llama-3-8B
Fine-tuning Method: LoRA (r=16, alpha=32)
Training: 3 epochs, batch_size=4, learning_rate=2e-4
Model Size: ~16 GB (FP16)
Parameters: 8 billion
Context Length: 8192 tokens
Language: Indonesian & English

Quick Start

Python (Transformers)

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Load model
model = AutoModelForCausalLM.from_pretrained(
    "emylton/arogya-ai-full",
    torch_dtype=torch.float16,
    device_map="auto"
)

tokenizer = AutoTokenizer.from_pretrained("emylton/arogya-ai-full")

# Generate prediction
prompt = """Prediksi kasus DBD di Kei Kecil untuk bulan depan berdasarkan data:
- Bulan ini: 45 kasus
- Bulan lalu: 38 kasus
- Curah hujan: tinggi
- Musim: penghujan"""

inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(
    **inputs,
    max_new_tokens=200,
    temperature=0.7,
    top_p=0.9,
    do_sample=True
)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Ollama

# 1. Download model
huggingface-cli download emylton/arogya-ai-full --local-dir ./arogya-full

# 2. Create Modelfile
cat > Modelfile << EOF
FROM ./arogya-full
PARAMETER temperature 0.7
PARAMETER top_p 0.9
PARAMETER num_ctx 8192
SYSTEM "Saya Arogya AI, asisten kesehatan untuk Kabupaten Maluku Tenggara. Saya dapat membantu analisis data kesehatan, prediksi penyakit, dan rekomendasi intervensi untuk 7 penyakit utama: DBD, ISPA, Malaria, Diare, TB, Stunting, dan Pneumonia."
EOF

# 3. Import to Ollama
ollama create arogya-ai -f Modelfile

# 4. Run
ollama run arogya-ai "Prediksi kasus DBD di Kei Kecil"

Use Cases

1. Disease Prediction

prompt = "Prediksi kasus Malaria di Kei Besar bulan Maret 2026"

2. Trend Analysis

prompt = "Analisis trend kasus ISPA di Maluku Tenggara 2023-2024"

3. Intervention Recommendations

prompt = "Rekomendasi intervensi untuk menurunkan kasus Stunting di Kei Kecil"

4. Resource Allocation

prompt = "Alokasi sumber daya kesehatan untuk program TB di 9 kecamatan"

5. Risk Assessment

prompt = "Penilaian risiko outbreak DBD di musim hujan"

Supported Diseases

DBD (Demam Berdarah Dengue)
ISPA (Infeksi Saluran Pernapasan Akut)
Malaria
Diare
TB (Tuberkulosis)
Stunting
Pneumonia

Geographic Coverage

9 Kecamatan di Kabupaten Maluku Tenggara:

Kei Kecil
Kei Besar
Kei Besar Selatan
Kei Besar Utara Timur
Kei Besar Utara Barat
Hoat Sorbay
Manyeuw
Kei Kecil Timur
Kei Kecil Barat

Training Details

Hyperparameters

{
    "lora_r": 16,
    "lora_alpha": 32,
    "lora_dropout": 0.05,
    "learning_rate": 2e-4,
    "num_train_epochs": 3,
    "per_device_train_batch_size": 4,
    "gradient_accumulation_steps": 4,
    "warmup_steps": 100,
    "max_seq_length": 2048,
    "optimizer": "paged_adamw_8bit"
}

Training Infrastructure

Platform: Google Colab Pro
GPU: NVIDIA A100 (40GB)
Training Time: ~6 hours
Framework: Transformers + PEFT + bitsandbytes

Data Processing

Data diproses melalui pipeline:

Excel Extraction: Pandas untuk structured data
PDF Extraction: PyMuPDF + tabula untuk tables
Data Cleaning: Normalisasi, deduplication
Prompt Engineering: Template khusus untuk health data
Train/Val Split: 90/10

Limitations

⚠️ Important Limitations:

Geographic Specificity: Model dilatih khusus untuk Maluku Tenggara, mungkin kurang akurat untuk daerah lain
Disease Coverage: Hanya 7 penyakit utama, tidak mencakup semua kondisi kesehatan
Data Timeframe: Data training dari 2020-2024, prediksi jangka panjang mungkin kurang akurat
Not Medical Advice: Model ini untuk analisis data, bukan pengganti konsultasi medis profesional
Language: Optimal untuk Bahasa Indonesia, kemampuan bahasa lain terbatas
Hallucination: Seperti LLM lainnya, model dapat menghasilkan informasi yang tidak akurat

Ethical Considerations

Intended Use

✅ Recommended:

Analisis data kesehatan populasi
Perencanaan program kesehatan
Alokasi sumber daya
Penelitian epidemiologi
Edukasi kesehatan masyarakat

❌ Not Recommended:

Diagnosis medis individual
Keputusan klinis tanpa verifikasi profesional
Pengganti tenaga kesehatan
Situasi darurat medis

Privacy & Security

Model tidak menyimpan data personal
Tidak ada identitas pasien dalam training data
Semua data diagregasi di level populasi
Ikuti regulasi kesehatan lokal saat menggunakan

Performance

Evaluation Metrics

Model dievaluasi pada validation set (10% dari data):

Perplexity: 2.34
Loss: 0.85
Accuracy (classification tasks): 87.3%
F1 Score: 0.86

Comparison

Model	Size	Accuracy	Use Case
Base Llama 3 8B	16 GB	45.2%	General
Arogya (LoRA)	27 MB	87.3%	Health (need base)
Arogya (Full)	16 GB	87.3%	Health (standalone)

Model Versions

Full Model vs LoRA Adapter

Aspect	LoRA Adapter	Full Model
Repository	emylton/arogya-health-model	emylton/arogya-ai-full
Size	27 MB	~16 GB
Download Time	1 min	10-30 min
Requires Base Model	✅ Yes (Llama 3 8B)	❌ No
Ollama Compatible	❌ No	✅ Yes
Best For	Developers/Researchers	End Users

Recommendation:

Use LoRA adapter if you already have Llama 3 8B or want to experiment
Use Full model for production deployment or Ollama usage

Citation

If you use this model in your research, please cite:

@software{arogya_ai_2024,
  title = {Arogya AI: Fine-tuned Language Model for Health Data Analysis in Maluku Tenggara},
  author = {Rafael and Contributors},
  year = {2024},
  url = {https://huggingface.co/emylton/arogya-ai-full},
  note = {Based on Meta-Llama-3-8B}
}

Research Paper:

@article{arogya_paper_2024,
  title = {Arogya AI: Implementasi Large Language Model untuk Analisis dan Prediksi Data Kesehatan di Kabupaten Maluku Tenggara},
  author = {Rafael and Contributors},
  year = {2024},
  url = {https://github.com/Rafael2022-prog/arogya-ai}
}

License

This model is based on Meta-Llama-3-8B and follows the Llama 3 Community License.

Additional Terms:

Model dapat digunakan untuk tujuan penelitian dan komersial
Wajib mencantumkan attribution
Tidak untuk tujuan yang merugikan atau melanggar hukum
Ikuti regulasi kesehatan setempat

Acknowledgments

Meta AI untuk Llama 3 base model
Dinas Kesehatan Kabupaten Maluku Tenggara untuk data kesehatan
Hugging Face untuk platform dan tools
Google Colab untuk training infrastructure

Contact & Support

GitHub Issues: https://github.com/Rafael2022-prog/arogya-ai/issues
Model Repository: https://huggingface.co/emylton/arogya-ai-full
Documentation: https://github.com/Rafael2022-prog/arogya-ai

Updates

Version 1.0 (2024)

Initial release
10,000+ training samples
7 diseases coverage
9 sub-districts coverage
Full model deployment

Built with ❤️ for better healthcare in Maluku Tenggara

Downloads last month: 29

Safetensors

Model size

8B params

Tensor type

F16

Model tree for emylton/arogya-ai-full

Base model

meta-llama/Meta-Llama-3-8B

Quantized

(276)

this model