Instructions to use fwizzer1/fwizzer-v3-ru with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use fwizzer1/fwizzer-v3-ru with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf fwizzer1/fwizzer-v3-ru:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf fwizzer1/fwizzer-v3-ru:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf fwizzer1/fwizzer-v3-ru:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf fwizzer1/fwizzer-v3-ru: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 fwizzer1/fwizzer-v3-ru:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf fwizzer1/fwizzer-v3-ru: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 fwizzer1/fwizzer-v3-ru:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf fwizzer1/fwizzer-v3-ru:Q4_K_M

Use Docker

docker model run hf.co/fwizzer1/fwizzer-v3-ru:Q4_K_M

LM Studio
Jan

vLLM

How to use fwizzer1/fwizzer-v3-ru with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "fwizzer1/fwizzer-v3-ru"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "fwizzer1/fwizzer-v3-ru",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/fwizzer1/fwizzer-v3-ru:Q4_K_M

Ollama
How to use fwizzer1/fwizzer-v3-ru with Ollama:
```
ollama run hf.co/fwizzer1/fwizzer-v3-ru:Q4_K_M
```
Unsloth Desktop

How to use fwizzer1/fwizzer-v3-ru with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf fwizzer1/fwizzer-v3-ru:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "fwizzer1/fwizzer-v3-ru:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use fwizzer1/fwizzer-v3-ru with Docker Model Runner:
```
docker model run hf.co/fwizzer1/fwizzer-v3-ru:Q4_K_M
```

Lemonade

How to use fwizzer1/fwizzer-v3-ru with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull fwizzer1/fwizzer-v3-ru:Q4_K_M

Run and chat with the model

lemonade run user.fwizzer-v3-ru-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use fwizzer1/fwizzer-v3-ru with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf fwizzer1/fwizzer-v3-ru: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 fwizzer1/fwizzer-v3-ru:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use fwizzer1/fwizzer-v3-ru with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf fwizzer1/fwizzer-v3-ru:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "fwizzer1/fwizzer-v3-ru:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

🚀 fwizzer-v3-ru (Fwizzer v3 Titan RU)

Флагманская русскоязычная рассуждающая модель v3
Базовая архитектура: Hybrid Liquid Neural Network (2.69B LFM2 Backbone)
Обучающий датасет: Приватный CoT-датасет fwizzer1/fwizzer-v3-titan-agentic (27 051 сэмплов с Kimi K2.5 CoT + 51 игра HTML5/WebGL)
Контекстное окно: до 32 768 токенов
Мышление: Глубокий CoT в тегах <think> ... </think> с архитектурным планом и самопроверкой

📦 Доступные сборки GGUF / Available Files

Файл	Размер	Назначение / Квант	Описание
`fwizzer-v3-ru-Max.gguf`	~3.4 ГБ	💎 Max (Q8_0)	Максимальная точность, нулевая деградация логики CoT
`fwizzer-v3-ru-Balanced.gguf`	~2.5 ГБ	⚖️ Balanced (Q5_K_M)	Золотой стандарт скорости и интеллекта для LM Studio / Ollama
`fwizzer-v3-ru-Speed.gguf`	~2.1 ГБ	⚡ Speed (Q4_K_M)	Сверхбыстрая легковесная версия для мобильных устройств и слабых GPU
`fwizzer-v3-ru-Speed-Q4_0.gguf`	~1.9 ГБ	🚀 Speed (Q4_0)	Классический компактный квант для запуска на любом CPU/GPU

🖥️ Быстрый запуск в LM Studio / Ollama

Скачайте квант (например, fwizzer-v3-ru-Balanced.gguf).
Поместите файл в папку моделей LM Studio или Ollama.
Запустите модель.

Системный промпт (System Prompt):

Ты думающая нейросеть а зовут тебя Fwizzer-R1-3B-RU. Весь ход мыслей и шаги пиши внутри тегов <think>(напиши сначала) и </think>(напиши по окончанию рассуждений), а итоговый ответ — обязательно после них.

Параметры сэмплинга:

Temperature: 0.6
Top-P: 0.95
Min-P: 0.05
Repetition Penalty: 1.05
Stop Tokens: </think>, <|im_end|>

🔬 Ключевые возможности Fwizzer v3:

Глубокий Chain-of-Thought: Формирование архитектурного плана и самопроверка перед выдачей ответа.
Генерация автономных игр и приложений: Создание 2D/3D игр на HTML5/Three.js в один файл без внешних зависимостей с управлением на WASD.
Процедурная генерация: Нативная поддержка GDScript (Godot 4), Python (Blender 4.x), Three.js WebGL.
Чистый процессор мышления: Отсутствие морализаторства и шаблонных отказов.

Downloads last month: 104

GGUF

Model size

3B params

Architecture

lfm2

Hardware compatibility

4-bit

View +3 variants

Model tree for fwizzer1/fwizzer-v3-ru

Base model

LiquidAI/LFM2.5-2.6B-Base

Finetuned

LiquidAI/LFM2.5-VL-3B

Quantized

(30)

this model