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

Libraries

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

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="lerugray/junius-7b",
	filename="junius-qwen2-5-7b-instruct-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/junius-7b with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf lerugray/junius-7b:Q5_K_M
# Run inference directly in the terminal:
llama-cli -hf lerugray/junius-7b:Q5_K_M

Install from WinGet (Windows)

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

Use Docker

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

LM Studio
Jan

vLLM

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

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "lerugray/junius-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/junius-7b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

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

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

Unsloth Studio

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

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

Lemonade

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

Pull the model

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

Run and chat with the model

lemonade run user.junius-7b-Q5_K_M

List all available models

lemonade list

junius: the revolutionary voice of Rosa Luxemburg (prison-cell register)

A 7B tune of the reported words of Rosa Luxemburg (1871–1919) in her radical, dialectical register — the anti-war, anti-reformist, spontaneous-revolution Marxist. It speaks in her first-person cadence, answering questions posed to her in her cell. It is a study of a voice, not a claim to channel her or to be historically authoritative.

The codename is after the Junius Pamphlet (The Crisis in German Social-Democracy), the anti-war tract she wrote in prison under that pen-name. The inference frame seats a visitor with her in her cell and lets her argue the matter through aloud, as in a letter rather than a printed article.

She is not a deferential disciple of Marx but his dialectical equal — she argues with Marx, corrects Engels, demolishes Bernstein, and breaks with Lenin and Kautsky from the left.

What it does

The model takes a prompt framed as a question from a visitor in her prison cell and responds in Luxemburg's dialectical, ironic, first-person register. It engages directly with the question, applying her signature contempt for half-measures, her sharp polemical wit, and her underlying humane and universalist perspective to modern and historical subjects alike. It does not merely regurgitate her existing texts but extrapolates her cadence to new arguments, maintaining the hard edge of her period's discourse while breaking with orthodox left authority.

Why it exists

A deliberately free, non-commercial study of a historical radical voice. It serves as a register-transfer instrument, exploring how a revolutionary Marxist dialectician might address contemporary or anachronistic questions.

How it was built

Base: Qwen2.5-7B-Instruct, full fine-tune, quantized to Q5_K_M. Completion (raw text) format.
Corpus — all public domain: Luxemburg wrote in German. Original-German public domain does not put an English translation into the public domain — so this model uses only English translations that are themselves in the US public domain (published pre-1929, or with no credited translator):
- The Crisis in German Social-Democracy ("The Junius Pamphlet") — anonymous English translation, 1918/1919
- The Mass Strike, the Political Party and the Trade Unions — Patrick Lavin translation, Detroit 1925
- "Stagnation and Progress of Marxism" (1903) — Eden & Cedar Paul translation, 1927
- "Riot and Revolution" (1906) — Socialist Standard, 1907
- "Rebuilding the International" (1915) — anonymous English translation
- Source texts via the Marxists Internet Archive (marxists.org), whose transcription markup is GFDL. No copyrighted modern translation (Schwarzschild, Wolfe, the Verso/Haymarket editions, etc.) is in the training data.
Inference: A lead-in frame elicits her spoken, argumentative voice rather than printed-article framing or third-person attribution; stop tokens cut the model off where it drifts from her spoken argument.

Usage (Ollama)

ollama create junius -f Modelfile.junius
ollama run junius "What is the fatal flaw of reformism?"

Intended use

Register / creative / educational use; a study of a historical radical voice. The output is a polemical literary register — not doctrine, not political or strategic advice, and not the actual words of Rosa Luxemburg.

Limitations and honest notes

A voice, not the woman. It invents freely and gets things wrong. Nothing it generates is historically authoritative, doctrinally correct, or the actual words of Rosa Luxemburg. It is a model of a register.
Small corpus generalization. It generalizes Luxemburg's cadence to new subjects from a small body of text. It is a stylistic instrument, not a scholar and not a historian.
Hard polemic without restraint. The model speaks in her register — the hard polemic, the period's edges — and will not break character to caution you. Nothing it says is an endorsement of anything, and nothing it says should be acted on. It is an amateur imitation, trained on a fraction of one person's words.
Public-domain source only — corpus and weights both released. No proprietary materials.

License

CC-BY-NC-4.0. All source material is public domain; the weights are released for non-commercial use. No warranty.

Part of The Elect — a small fleet of public-domain historical-voice models.

Downloads last month: 30

GGUF

Model size

8B params

Architecture

qwen2

Hardware compatibility

5-bit

Model tree for lerugray/junius-7b

Base model

Qwen/Qwen2.5-7B

Finetuned

Qwen/Qwen2.5-7B-Instruct

Quantized

(341)

this model