Instructions to use NoirZangetsu/Qwen2.5-3B-Flutter with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NoirZangetsu/Qwen2.5-3B-Flutter with Transformers:

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

pipe = pipeline("text-generation", model="NoirZangetsu/Qwen2.5-3B-Flutter")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("NoirZangetsu/Qwen2.5-3B-Flutter", dtype="auto")

llama-cpp-python

How to use NoirZangetsu/Qwen2.5-3B-Flutter with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="NoirZangetsu/Qwen2.5-3B-Flutter",
	filename="unsloth.Q4_K_M.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use NoirZangetsu/Qwen2.5-3B-Flutter with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf NoirZangetsu/Qwen2.5-3B-Flutter: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 NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf NoirZangetsu/Qwen2.5-3B-Flutter: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 NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M

Use Docker

docker model run hf.co/NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M

LM Studio
Jan

vLLM

How to use NoirZangetsu/Qwen2.5-3B-Flutter with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NoirZangetsu/Qwen2.5-3B-Flutter"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NoirZangetsu/Qwen2.5-3B-Flutter",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M

SGLang

How to use NoirZangetsu/Qwen2.5-3B-Flutter 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 "NoirZangetsu/Qwen2.5-3B-Flutter" \
    --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": "NoirZangetsu/Qwen2.5-3B-Flutter",
		"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 "NoirZangetsu/Qwen2.5-3B-Flutter" \
        --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": "NoirZangetsu/Qwen2.5-3B-Flutter",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use NoirZangetsu/Qwen2.5-3B-Flutter with Ollama:
```
ollama run hf.co/NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M
```

Unsloth Studio new

How to use NoirZangetsu/Qwen2.5-3B-Flutter 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 NoirZangetsu/Qwen2.5-3B-Flutter 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 NoirZangetsu/Qwen2.5-3B-Flutter to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for NoirZangetsu/Qwen2.5-3B-Flutter to start chatting

Pi new

How to use NoirZangetsu/Qwen2.5-3B-Flutter with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf NoirZangetsu/Qwen2.5-3B-Flutter: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": "NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use NoirZangetsu/Qwen2.5-3B-Flutter with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf NoirZangetsu/Qwen2.5-3B-Flutter: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 NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use NoirZangetsu/Qwen2.5-3B-Flutter with Docker Model Runner:
```
docker model run hf.co/NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M
```

Lemonade

How to use NoirZangetsu/Qwen2.5-3B-Flutter with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull NoirZangetsu/Qwen2.5-3B-Flutter:Q4_K_M

Run and chat with the model

lemonade run user.Qwen2.5-3B-Flutter-Q4_K_M

List all available models

lemonade list

🦋 Qwen2.5-3B-Flutter — Türkçe Flutter Kod Anlama ve Üretme Modeli

🚀 Fine-tuned by NoirZangetsu using Unsloth and 🤗 Hugging Face TRL library.
💡 Amacı: Türkçe yazılmış Flutter soru-cevap verisiyle eğitilmiş, kod üretme ve anlama görevlerinde özelleştirilmiş bir Qwen2.5 3B modeli.

🧠 Model Bilgileri

Özellik	Detay
Model	`Qwen2.5-3B` (Unsloth optimized, 4-bit)
Parametre Sayısı	3.09B
Eğitim Kitaplığı	`Unsloth`, `Hugging Face Transformers`, `TRL`
Lisans	`Apache 2.0`
Donanım	GTX 1650 Mobile (4GB VRAM), 16GB RAM
Eğitim Süresi	~1.5 saat (QLoRA ile 4-bit hızlı fine-tune)
Dataset Boyutu	`990 örnek` (csv formatında)
Dataset Türü	Kodlama / Soru-Cevap (Türkçe, Flutter özelinde)

📚 Kullanılan Veri Seti

Model, Flutter ile ilgili Türkçe yazılmış soru ve cevaplardan oluşan özel bir veri seti ile eğitildi:

Veri Kümesi Adı: flutter_code_with_questions_990.csv
İçerik Özeti:

Türkçe yazılmış Flutter ile ilgili teknik sorular
Kod blokları, açıklamalar ve çözüm adımları
Türk yazılımcılar için lokal bağlamda anlamlı içerikler

📊 Performans Karşılaştırması

Aşağıdaki grafik, modelin 3 farklı metrik üzerinden kıyaslamasını gösterir:

Model	BLEU	ROUGE-L	METEOR
Qwen2.5-3B-Flutter	0.06	0.14	0.14
Qwen2.5-Coder-3B-Instruct	0.01	0.09	0.05

🔍 Yorum:

Flutter'a özel veri ile fine-tune edilen model, özellikle ROUGE-L ve METEOR skorlarında ciddi bir iyileşme göstermiştir. Bu da modelin hem semantik benzerliği hem de anlamlı parça üretimi konusunda daha başarılı olduğunu gösteriyor.

🛠️ Kullanım

from transformers import AutoTokenizer, AutoModelForCausalLM

model = AutoModelForCausalLM.from_pretrained("NoirZangetsu/Qwen2.5-3B-Flutter", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("NoirZangetsu/Qwen2.5-3B-Flutter")

prompt = "Flutter'da bir StatefulWidget nasıl tanımlanır?"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

🧪 Değerlendirme Metodolojisi

Model aşağıdaki metrikler kullanılarak değerlendirildi:

BLEU: Cümle düzeyinde n-gram eşleşmeleri
ROUGE-L: En uzun ortak alt dizi üzerinden ölçüm
METEOR: Levenshtein benzerliği + semantik eşleşmeler

📌 Ekstra Notlar

Model 4-bit formatta fine-tune edilmiştir. Bellek kullanımı açısından son derece verimlidir.
Eğitim verisi küçük ama yüksek kaliteli olduğundan, model transfer öğrenme açısından efektif sonuç vermiştir.
Daha büyük veri setleri ile yeniden fine-tune edilerek çok daha iyi skorlar elde edilebilir.

🧠 Gelecek Planları

Daha büyük veri setiyle tekrar fine-tune
Türkçe test benchmark setleriyle karşılaştırmalı analiz
Web arayüzü üzerinden demo yayını

🤝 Katkıda Bulunmak

Modeli test ettiyseniz, çıktı örneklerini veya önerilerinizi paylaşarak gelişime katkı sunabilirsiniz. 🙌

📤 Citation

@misc{qwen25flutter2025,
  title={Qwen2.5-3B-Flutter: Türkçe Flutter Kod Anlama ve Üretme Modeli},
  author={NoirZangetsu},
  year={2025},
  howpublished={\url{https://huggingface.co/NoirZangetsu/Qwen2.5-3B-Flutter}},
}

Downloads last month: 147

GGUF

Model size

3B params

Architecture

qwen2

Hardware compatibility

4-bit

Model tree for NoirZangetsu/Qwen2.5-3B-Flutter

Base model

Qwen/Qwen2.5-3B

Finetuned

Qwen/Qwen2.5-Coder-3B

Finetuned

Qwen/Qwen2.5-Coder-3B-Instruct

Quantized

(98)

this model