Instructions to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with Transformers:

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

pipe = pipeline("text-generation", model="wololoo/Llama-3.2-3B-TR-Instruct-DPO")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("wololoo/Llama-3.2-3B-TR-Instruct-DPO")
model = AutoModelForCausalLM.from_pretrained("wololoo/Llama-3.2-3B-TR-Instruct-DPO")
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]:]))

PEFT
How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with PEFT:
```
Task type is invalid.
```

llama-cpp-python

How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="wololoo/Llama-3.2-3B-TR-Instruct-DPO",
	filename="Llama-3.2-3B-Tr-Muhendis-Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Inference
Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_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 wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_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 wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M

Use Docker

docker model run hf.co/wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M

LM Studio
Jan

vLLM

How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "wololoo/Llama-3.2-3B-TR-Instruct-DPO"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "wololoo/Llama-3.2-3B-TR-Instruct-DPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M

SGLang

How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO 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 "wololoo/Llama-3.2-3B-TR-Instruct-DPO" \
    --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": "wololoo/Llama-3.2-3B-TR-Instruct-DPO",
		"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 "wololoo/Llama-3.2-3B-TR-Instruct-DPO" \
        --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": "wololoo/Llama-3.2-3B-TR-Instruct-DPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with Ollama:
```
ollama run hf.co/wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M
```

Unsloth Studio

How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO 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 wololoo/Llama-3.2-3B-TR-Instruct-DPO 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 wololoo/Llama-3.2-3B-TR-Instruct-DPO to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for wololoo/Llama-3.2-3B-TR-Instruct-DPO to start chatting

How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with Docker Model Runner:
```
docker model run hf.co/wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M
```

Lemonade

How to use wololoo/Llama-3.2-3B-TR-Instruct-DPO with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull wololoo/Llama-3.2-3B-TR-Instruct-DPO:Q4_K_M

Run and chat with the model

lemonade run user.Llama-3.2-3B-TR-Instruct-DPO-Q4_K_M

List all available models

lemonade list

Llama-3.2-3B-TR-Instruct-DPO

Bu model, Llama-3.2-3B-Instruct temel alınarak geliştirilmiş, Türkçe dil yetenekleri ve teknik (STEM) alanlardaki muhakeme gücü artırılmış bir yapay zeka asistanıdır.

This model is a Turkish fine-tuned version of unsloth/Llama-3.2-3B-Instruct. It has been trained using a two-stage pipeline (SFT + DPO) to enhance its Turkish language capabilities and reasoning skills in STEM (Science, Technology, Engineering, Mathematics) fields.

Model Details (Model Detayları)

Model Description

Bu model, talimatları daha iyi takip edebilmesi ve insan tercihlerine uygun cevaplar verebilmesi için iki aşamalı bir eğitimden geçmiştir:

SFT (Supervised Fine-Tuning): atasoglu/databricks-dolly-15k-tr veri seti kullanılarak genel konuşma ve talimat takip yeteneği kazandırıldı.
DPO (Direct Preference Optimization): yusufbaykaloglu/Turkish-STEM-DPO-Dataset veri seti ile modelin özellikle bilim ve teknoloji konularındaki cevap kalitesi optimize edildi.

Developed by: wololoo
Model type: Causal Language Model (Transformer)
Language(s): Turkish (Türkçe)
License: Llama 3.2 Community License
Finetuned from model: unsloth/Llama-3.2-3B-Instruct

Uses (Kullanım Alanları)

Direct Use

Bu model genel amaçlı bir Türkçe asistan olarak kullanılabilir. Özellikle şu alanlarda etkilidir:

Türkçe soru-cevap (QA).
Bilim, Teknoloji ve Mühendislik (STEM) konularında temel bilgilendirme.
Metin üretimi, özetleme ve düzenleme.
Sohbet (Chatbot) uygulamaları.

Out-of-Scope Use (Kapsam Dışı Kullanım)

Model tıbbi, hukuki veya finansal tavsiye vermek için kullanılmamalıdır.
Zararlı, nefret söylemi içeren veya yasa dışı içerik üretimi için kullanılamaz.

Bias, Risks, and Limitations (Riskler ve Limitler)

Tüm Büyük Dil Modelleri (LLM) gibi, bu model de "halüsinasyon" görebilir (yanlış bilgi üretebilir) veya eğitim verisindeki önyargıları yansıtabilir. Özellikle teknik veya kritik konularda modelin verdiği cevaplar mutlaka doğrulanmalıdır.

How to Get Started with the Model (Nasıl Kullanılır?)

Modeli çalıştırmak için aşağıdaki Python kodunu kullanabilirsiniz:

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

# Model ismini buraya girin
model_id = "wololoo/Llama-3.2-3B-TR-Instruct-DPO"

# Tokenizer ve Modeli Yükle
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.bfloat16,
    device_map="auto"
)

# Sohbet Şablonunu Hazırla
messages = [
    {"role": "system", "content": "Sen yardımsever ve bilgili bir yapay zeka asistanısın."},
    {"role": "user", "content": "Yapay zeka mühendisliğinde DPO (Direct Preference Optimization) nedir?"},
]

input_ids = tokenizer.apply_chat_template(
    messages,
    add_generation_prompt=True,
    return_tensors="pt"
).to(model.device)

# Cevap Üret
outputs = model.generate(
    input_ids,
    max_new_tokens=256,
    do_sample=True,
    temperature=0.7,
    top_p=0.9,
)

# Sonucu Yazdır
response = tokenizer.decode(outputs[0][input_ids.shape[-1]:], skip_special_tokens=True)
print(response)

Downloads last month: 115

Safetensors

Model size

3B params

Tensor type

F16

Model tree for wololoo/Llama-3.2-3B-TR-Instruct-DPO

Base model

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

Finetuned

unsloth/Llama-3.2-3B-Instruct

Quantized

(111)

this model

wololoo
/

Llama-3.2-3B-TR-Instruct-DPO