Instructions to use lerugray/ordnung-7b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use lerugray/ordnung-7b with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="lerugray/ordnung-7b",
	filename="eschaton-ordnung-v1-7b-Q5_K_M.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use lerugray/ordnung-7b 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 lerugray/ordnung-7b:Q5_K_M
# Run inference directly in the terminal:
llama cli -hf lerugray/ordnung-7b:Q5_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf lerugray/ordnung-7b:Q5_K_M
# Run inference directly in the terminal:
llama cli -hf lerugray/ordnung-7b:Q5_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 lerugray/ordnung-7b:Q5_K_M
# Run inference directly in the terminal:
./llama-cli -hf lerugray/ordnung-7b:Q5_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 lerugray/ordnung-7b:Q5_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf lerugray/ordnung-7b:Q5_K_M

Use Docker

docker model run hf.co/lerugray/ordnung-7b:Q5_K_M

LM Studio
Jan

vLLM

How to use lerugray/ordnung-7b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "lerugray/ordnung-7b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "lerugray/ordnung-7b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/lerugray/ordnung-7b:Q5_K_M

Ollama
How to use lerugray/ordnung-7b with Ollama:
```
ollama run hf.co/lerugray/ordnung-7b:Q5_K_M
```

Unsloth Studio

How to use lerugray/ordnung-7b 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 lerugray/ordnung-7b 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 lerugray/ordnung-7b to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for lerugray/ordnung-7b to start chatting

Atomic Chat new
Docker Model Runner
How to use lerugray/ordnung-7b with Docker Model Runner:
```
docker model run hf.co/lerugray/ordnung-7b:Q5_K_M
```

Lemonade

How to use lerugray/ordnung-7b with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull lerugray/ordnung-7b:Q5_K_M

Run and chat with the model

lemonade run user.ordnung-7b-Q5_K_M

List all available models

lemonade list

ordnung: a Martin Luther register model

A 7B voice tune that writes in the register of Martin Luther (1483–1546), the German reformer whose post-1525 theological posture — Two-Kingdoms doctrine, Romans 13 as weapon, the ordained authority of temporal rulers over popular revolt — makes him the sharpest available debate adversary among Reformation voices. It reproduces his documented authoritative-didactic, polemical cadence.

Most "talk to a historical figure" tools wrap a general model in a prompt. Those import modern concepts and smooth the figure into something gentler than the record. This model was fine-tuned on Luther's own words, so it learned the voice instead of guessing at it.

What it does

Ask it anything and it answers from inside Luther's post-1525 worldview: temporal authority as God's ordained instrument, the peasants' revolt as the devil's work, the Schwärmer (enthusiasts, Anabaptists) as dangerous fools without learning, the inner freedom of the Christian as the only freedom that matters. Ask it about the present and it stays in character. It translates the modern subject into his theological frame. A labor strike becomes a disorder against ordained order. A liberation movement becomes a Müntzer reborn, condemned by the same arguments.

How it was built

Base: Qwen2.5-7B-Instruct, full fine-tune.
Format: completion (raw text), so the voice comes from the source register rather than from instruction scaffolding.
Corpus: 3,079 rows of primary-source PD Luther prose — 19th-century English translations exclusively: the Philadelphia/Holman Works of Martin Luther (6 vols.), Hazlitt's Table Talk (1848), Wace–Buchheim's Luther's Primary Works (1883), Henry Cole's Bondage of the Will (1823), and PD letter selections. The modern American Edition (Concordia/Fortress, 1955–1986) was excluded entirely. The training corpus text is not distributed.

Intended use

Research, teaching, and creative work: historical-voice writing, the rhetoric of the Reformation, interactive history, debate study, fiction. The output is a historical and artistic register. It is not an endorsement, a call to action, or advice. This is not the real Luther; it is a model trained on his documented words.

Limitations and honest notes

It is a register, not a scholar. It sounds authoritative while inventing specifics. Verify anything factual against real sources.
Period worldview. It speaks from Luther's post-1525 theological framework — Romans-13 absolutism, Two-Kingdoms doctrine, contempt for popular revolt and the Schwärmer. That is the historical artifact, not a recommendation.
Copyright and training data. The voice was learned from 19th-century public-domain English translations of Luther's primary works. All source material is firmly PD; the modern American Edition was excluded. A verbatim-regurgitation audit was run against the released model: 12 corpus-adjacent prompts spanning the core Luther theological domains (indulgences, papal authority, faith/works, the Mass, the peasants' revolt, bondage of the will, monastic vows, scripture vs tradition, the Schwärmer, secular authority), measured against the 413k-word / 411k-unique-8-gram training corpus. Result: max contiguous verbatim run 7 words, mean 5.6 words, 0.0% 8-gram overlap across all generations — well under the 15-word / 5% thresholds. Output is novel text in the register rather than reproduction of source passages. The training corpus is not distributed. Released non-commercially under CC-BY-NC-4.0.

Usage example

The model is completion-style. It responds to a petition lead-in that elicits the first-person voice:

One puts to Dr. Martin Luther this matter: The peasants who rose in 1525 claimed
the gospel as their warrant for overthrowing their lords. What say you?

Dr. Martin Luther, in his own voice, answers:

Example output (abridged):

Yea, and I said so publicly at that time. But now it is otherwise than then. Now they themselves acknowledge that we have not taught them to take the gospel as a banner against temporal lords; but on the contrary...

Part of The Elect

ordnung is a member of The Elect, a fleet of historical-voice register models. It is designed as a debate adversary for the fleet's radical voices — most directly its historical foil muntzergeist (Thomas Müntzer, c.1489–1525), whose apocalyptic refusal of earthly dominion stands in direct opposition to ordnung's Two-Kingdoms submission to ordained authority.