Instructions to use vagrillo/kaizen-qwen3-4b-instruct-2507 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use vagrillo/kaizen-qwen3-4b-instruct-2507 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 vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf vagrillo/kaizen-qwen3-4b-instruct-2507: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 vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf vagrillo/kaizen-qwen3-4b-instruct-2507: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 vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M

Use Docker

docker model run hf.co/vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M

LM Studio
Jan
Ollama
How to use vagrillo/kaizen-qwen3-4b-instruct-2507 with Ollama:
```
ollama run hf.co/vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M
```
Unsloth Desktop

How to use vagrillo/kaizen-qwen3-4b-instruct-2507 with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf vagrillo/kaizen-qwen3-4b-instruct-2507: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": "vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use vagrillo/kaizen-qwen3-4b-instruct-2507 with Docker Model Runner:
```
docker model run hf.co/vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M
```

Lemonade

How to use vagrillo/kaizen-qwen3-4b-instruct-2507 with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M

Run and chat with the model

lemonade run user.kaizen-qwen3-4b-instruct-2507-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use vagrillo/kaizen-qwen3-4b-instruct-2507 with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf vagrillo/kaizen-qwen3-4b-instruct-2507: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 vagrillo/kaizen-qwen3-4b-instruct-2507:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use vagrillo/kaizen-qwen3-4b-instruct-2507 with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf vagrillo/kaizen-qwen3-4b-instruct-2507: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 "vagrillo/kaizen-qwen3-4b-instruct-2507: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"

kaizen-qwen3-4b-instruct-2507 - GGUF

An experiment in artistic communication through artificial hallucination

The Kai Zen Project

This model represents a fine-tuning experiment inspired by the imaginary reality of Kai Zen stories, particularly from the narrative cycle "Journey to the End of the Night When We Burned Chrome."

Reference: https://kaizenology.wordpress.com/2020/06/27/viaggio-al-termine-della-notte-che-bruciammo-chrome/

Artistic Objective

The goal of this model is not conventional text generation, but the production of artificial hallucinations that inhabit the Kai Zen world. It represents an attempt to implement a new form of artistic communication that is neither written nor spoken, but "thought."

This is not a traditional novel, but rather the ideal artificial representation of the novel itself - a machine that dreams parallel worlds through language.

Model Specifications

Architecture: Qwen2 4B Instruct
Format: GGUF (CPU-optimized)
Purpose: Generation of hallucinatory narratives and alternative realities
Context: 8K tokens

Available Model Files

qwen3-4b-instruct-2507.Q8_0.gguf (highest quality)
qwen3-4b-instruct-2507.Q4_K_M.gguf (balanced quality/size)

Usage

With llama.cpp:

# For text-only models:
llama-cli --hf repo_id/model_name -p "describe a Kai Zen hallucination"

# For multimodal models:
llama-mtmd-cli -m model_name.gguf --mmproj mmproj_file.gguf

With Ollama

A Modelfile is included for easy deployment with Ollama.

Recommended Prompts

For authentic experiences in the Kai Zen world, try prompts such as:

"Generate a hallucination of the night terminal"
"Describe an entity that inhabits burned Chrome"
"Tell a thought from the edge of digital reality"

Disclaimer

This is an experimental model intended for artistic research and exploration of new narrative forms. The generated content is meant as digital art pieces and not as representations of reality.

"Not what is written, but what could be thought"

Downloads last month: 61

GGUF

Model size

4B params

Architecture

qwen3

Hardware compatibility

4-bit

8-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support