Instructions to use roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf 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 roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf: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 roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf: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 roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M

Use Docker

docker model run hf.co/roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M

LM Studio
Jan
Ollama
How to use roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf with Ollama:
```
ollama run hf.co/roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M
```
Unsloth Desktop

How to use roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf: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": "roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf with Docker Model Runner:
```
docker model run hf.co/roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M
```

Lemonade

How to use roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M

Run and chat with the model

lemonade run user.gemma-4-E2B-it.Q4_K_M-med.gguf-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf: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 roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf: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 "roelfrenkema/gemma-4-E2B-it.Q4_K_M-med.gguf: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"

gemma-4-E2B-it.Q4_K_M-med.gguf / README

roelfrenkema

first commit

5806b76 6 months ago

Raw

History Blame Contribute Delete

1.35 kB

	# gemma-4-E2B-it.Q4_K_M-med.gguf

	**Dit model is een gespecialiseerd taalmodel, gefinetuned voor de analyse van medische research en psychiatrische diagnostiek.

	## 📊 Training Highlights
	Het model is getraind op (Debian 13) met gebruik van de Unsloth Studio en een NVIDIA RTX 3060 12GB GPU.

	* Base Model: `unsloth/gemma-4-E2B-it`
	* Methode: QLoRA (4-bit)
	* Dataset: Gespecialiseerde medische dataset roelfrenkema/medical_instruct_nl_full
	* Eind-Loss: 1.4544 (na 4242 stappen)
	* Stabiliteit: Grad Norm van 2.320, wat duidt op een zeer stabiel leerproces zonder significante hallucinaties
	* Training Time: ~6.5 uur op een RTX 3060

	## 🚀 Gebruik met Ollama
	Voor optimale resultaten en het voorkomen van de "Thinking Process" weergave en tekstherhalingen, gebruik de volgende instellingen in je `Modelfile`:

	Modelfile configuratie:
	```
	FROM ./gemma-4-E2B-it.Q4_K_M-med.gguf
	SYSTEM "Je bent Dr. Gemma, een arts. Beantwoord medische vragen uitsluitend op basis van de context. Toon GEEN intern denkproces (Thinking Process). Begin direct met het feitelijke Nederlandse antwoord."
	PARAMETER temperature 0.3
	PARAMETER repeat_penalty 1.6
	PARAMETER repeat_last_n 128
	PARAMETER top_k 40
	PARAMETER top_p 0.9
	```

	---
	Ontwikkeld door: Roelf Renkema & Gemi (AI Partner)
	Datum: 4 april 2026
	Licentie: Apache