Instructions to use Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor 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 Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0
# Run inference directly in the terminal:
llama cli -hf Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0
# Run inference directly in the terminal:
llama cli -hf Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

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 Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0
# Run inference directly in the terminal:
./llama-cli -hf Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

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 Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

Use Docker

docker model run hf.co/Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

LM Studio
Jan
Ollama
How to use Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor with Ollama:
```
ollama run hf.co/Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0
```
Unsloth Desktop

How to use Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

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": "Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor with Docker Model Runner:
```
docker model run hf.co/Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0
```

Lemonade

How to use Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

Run and chat with the model

lemonade run user.Gemma-4-26B-A4B-StylizedDecensor-Q4_0

List all available models

lemonade list

Hermes Agent

How to use Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

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 Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0

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 "Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor:Q4_0" \
  --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"

Gemma-4-26B-A4B-StylizedDecensor

An experimental, heavily modified GGUF merge combining QAT-quantized uncensored weights with high-precision StyleTune layers for the Gemma-4 26B MoE architecture.

Status: Experimental / Unstable. Use with caution.

Model Composition & Tensor Surgery

This model utilizes direct GGUF tensor surgery:

Base Model: OS-Software's QAT-quantized uncensored Heretic v2 (gemma-4-26B-A4B-it-qat-q4_0-uncensored-heretic-v2-GGUF).
Tensor Swap: Because this architecture uses tied embeddings, the token_embd.weight tensor (which acts as both the input embedding and the output.weight lm_head) was surgically replaced using the higher-precision Q6_K version from mradermacher's StyleTune V2 (Gemma-4-26B-A4B-StyleTune-V2.Q6_K.gguf).

This ensures the vocabulary projection space remains clean while preserving the underlying uncensored behavior.

Known Quirks & Behavior

High Creativity, Low Stability: The model is exceptionally creative with its vocabulary and phrasing, but it is inherently volatile. It can drift or break character more easily than standard fine-tunes.

Prompt Format

This model follows the standard Gemma-4 turn-wrapper formatting structure. Ensure your loader or client uses the correct template mapping.

Downloads last month: 166

GGUF

Model size

25B params

Architecture

gemma4

Hardware compatibility

4-bit

Inference Providers NEW

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

Model tree for Kanjdoes/Gemma-4-26B-A4B-StylizedDecensor

Base model

google/gemma-4-26B-A4B

Finetuned

google/gemma-4-26B-A4B-it

Finetuned

Gryphe/Gemma-4-26B-A4B-StyleTune-V2

Quantized

(14)

this model