Instructions to use HexQuant/Pars-Medical-o1-Llama-FFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use HexQuant/Pars-Medical-o1-Llama-FFT with Transformers:

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

pipe = pipeline("text-generation", model="HexQuant/Pars-Medical-o1-Llama-FFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("HexQuant/Pars-Medical-o1-Llama-FFT")
model = AutoModelForCausalLM.from_pretrained("HexQuant/Pars-Medical-o1-Llama-FFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use HexQuant/Pars-Medical-o1-Llama-FFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "HexQuant/Pars-Medical-o1-Llama-FFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HexQuant/Pars-Medical-o1-Llama-FFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/HexQuant/Pars-Medical-o1-Llama-FFT

SGLang

How to use HexQuant/Pars-Medical-o1-Llama-FFT 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 "HexQuant/Pars-Medical-o1-Llama-FFT" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HexQuant/Pars-Medical-o1-Llama-FFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "HexQuant/Pars-Medical-o1-Llama-FFT" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "HexQuant/Pars-Medical-o1-Llama-FFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Studio new

How to use HexQuant/Pars-Medical-o1-Llama-FFT 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 HexQuant/Pars-Medical-o1-Llama-FFT 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 HexQuant/Pars-Medical-o1-Llama-FFT to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for HexQuant/Pars-Medical-o1-Llama-FFT to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="HexQuant/Pars-Medical-o1-Llama-FFT",
    max_seq_length=2048,
)

Docker Model Runner
How to use HexQuant/Pars-Medical-o1-Llama-FFT with Docker Model Runner:
```
docker model run hf.co/HexQuant/Pars-Medical-o1-Llama-FFT
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

🩺 Pars-Medical-o1-Llama-FFT

⚠️This is a beta version and does not have the ability to correctly respond to or handle local runs on its own.⚠️

⚠️The full version will be released in the Pars-Medical application.⚠️

اولین مدل پزشکی دوزبانه فارسی-انگلیسی با استدلال زنجیره‌ای

First Bilingual Persian-English Medical Model with Chain-of-Thought Reasoning

🌟 Overview | معرفی

پارس-مدیکال یک مدل هوش مصنوعی پزشکی است که با روش Full Fine-Tune روی مجموعه داده‌های پزشکی فارسی و انگلیسی آموزش دیده است.

این مدل از تکنیک Chain-of-Thought (استدلال زنجیره‌ای) مشابه مدل‌های o1 استفاده می‌کند تا پاسخ‌های دقیق‌تر و منطقی‌تر ارائه دهد.

Pars-Medical is a medical AI model trained with Full Fine-Tuning on Persian and English medical datasets.

It uses Chain-of-Thought reasoning (similar to o1-style models) to provide more accurate and logical responses.

✨ Features | ویژگی‌ها

Feature	Description	توضیحات
🌐	Bilingual - Persian & English	دوزبانه - فارسی و انگلیسی
🧠	Chain-of-Thought - Step-by-step reasoning	استدلال زنجیره‌ای - گام به گام
🔬	Full Fine-Tune - Not LoRA/adapter	آموزش کامل - نه آداپتور
📊	48K+ Samples - Large medical dataset	+۴۸ هزار نمونه - دیتاست بزرگ
⚡	3B Parameters - Fast inference	۳ میلیارد پارامتر - استنتاج سریع
🏥	Clinical Focus - Diagnosis & treatment	تمرکز بالینی - تشخیص و درمان

🚀 Quick Start | شروع سریع

Using Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "HexQuant/Pars-Medical-o1-Llama-FFT"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
    device_map="auto"
)

messages = [
    {"role": "user", "content": "علائم دیابت نوع ۲ چیست؟"}
]

inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to(model.device)
outputs = model.generate(inputs, max_new_tokens=512, temperature=0.7)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

Using Ollama (GGUF)

# Use quantized version for local inference
ollama run hf.co/HexQuant/Pars-Medical-o1-Llama-FFT-GGUF

📦 Looking for GGUF? See Pars-Medical-o1-Llama-FFT-GGUF for quantized versions.

💬 Example | نمونه

🇮🇷 مثال فارسی - تشخیص پزشکی

User:

یک مرد ۵۵ ساله با درد قفسه سینه، تنگی نفس و تعریق مراجعه کرده است. علائم از ۲ ساعت پیش شروع شده. تشخیص احتمالی چیست؟

Model:

<think> بیایید این مورد را گام به گام بررسی کنیم:

۱. مشخصات بیمار: مرد ۵۵ ساله - در گروه سنی پرخطر برای بیماری‌های قلبی

۲. علائم کلیدی:

درد قفسه سینه (علامت کلاسیک)

تنگی نفس (نشانه اختلال عملکرد قلب)

تعریق (فعال شدن سیستم سمپاتیک)

۳. شروع حاد: ۲ ساعت پیش - نشان‌دهنده وضعیت اورژانسی

۴. تشخیص افتراقی:

سندرم حاد کرونری (ACS) ⚠️ محتمل‌ترین

آمبولی ریه

دیسکسیون آئورت </think>

تشخیص احتمالی: سندرم حاد کرونری (Acute Coronary Syndrome)

اقدامات فوری توصیه شده:

ECG فوری (در ۱۰ دقیقه اول)

تروپونین و آنزیم‌های قلبی

آسپرین ۳۲۵ میلی‌گرم

مشاوره قلب اورژانسی

🇺🇸 English Example - Clinical Reasoning

User:

A 45-year-old woman presents with fatigue, weight gain, cold intolerance, and constipation for 3 months. What is the likely diagnosis?

Model:

<think> Let me analyze this systematically:

Patient Profile: 45-year-old female - higher risk for thyroid disorders

Symptom Analysis:

Fatigue → decreased metabolic rate

Weight gain → slowed metabolism

Cold intolerance → reduced thermogenesis

Constipation → decreased GI motility

Pattern Recognition: All symptoms point to hypometabolic state

Most Likely Cause: These are classic signs of decreased thyroid hormone </think>

Likely Diagnosis: Hypothyroidism

Recommended workup:

TSH (primary screening test)

Free T4

Anti-TPO antibodies (if Hashimoto's suspected)

Lipid panel (dyslipidemia common in hypothyroidism)

📊 Training Details | جزئیات آموزش

Parameter	Value
Base Model	`Llama-3.2-3B-Instruct`
Method	Full Fine-Tune (FFT)
Framework	Unsloth + Transformers
Precision	BF16
Context Length	4096 tokens
Training Hardware	NVIDIA A100 80GB

Datasets | دیتاست‌ها

Dataset	Language	Samples
medical-o1-reasoning-SFT	English	~40,000
persian-medical-qa	Persian	~5,000
Persian-Medical-Dataset	Persian	~3,000

📦 Available Formats | فرمت‌های موجود

Format	Size	Use Case	Link
BF16 (This repo)	~6.4 GB	Fine-tuning, Full precision	Download
GGUF Q4_K_M	~2.0 GB	Ollama, llama.cpp, LM Studio	Download
GGUF Q5_K_M	~2.5 GB	Better quality, local use	Download
GGUF Q8_0	~3.5 GB	Best quality quantized	Download

⚠️ Disclaimer | سلب مسئولیت

⚠️ این مدل صرفاً برای اهداف آموزشی و تحقیقاتی طراحی شده است.

❌ جایگزین مشاوره پزشکی حرفه‌ای نیست.

✅ همیشه با پزشک متخصص مشورت کنید.

⚠️ This model is for educational and research purposes only.

❌ Not a substitute for professional medical advice.

✅ Always consult with a qualified healthcare provider.

🔗 Connect | ارتباط

Made with ❤️ TheRealPourya

If you find this model useful, please ⭐ star the repo and share it!

Downloads last month: 5

Safetensors

Model size

3B params

Tensor type

BF16

Model tree for HexQuant/Pars-Medical-o1-Llama-FFT

Base model

meta-llama/Llama-3.2-3B-Instruct

Finetuned

unsloth/Llama-3.2-3B-Instruct