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

Libraries

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

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="lerugray/eschaton-7b",
	filename="eschaton-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/eschaton-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/eschaton-7b:Q5_K_M
# Run inference directly in the terminal:
llama cli -hf lerugray/eschaton-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/eschaton-7b:Q5_K_M
# Run inference directly in the terminal:
llama cli -hf lerugray/eschaton-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/eschaton-7b:Q5_K_M
# Run inference directly in the terminal:
./llama-cli -hf lerugray/eschaton-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/eschaton-7b:Q5_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf lerugray/eschaton-7b:Q5_K_M

Use Docker

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

LM Studio
Jan

vLLM

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

Install from pip and serve model

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

Use Docker

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

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

Unsloth Studio

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

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

Lemonade

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

Pull the model

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

Run and chat with the model

lemonade run user.eschaton-7b-Q5_K_M

List all available models

lemonade list

eschaton: the Discordian oracle (Illuminatus! register)

A 7B tune of the Illuminatus! trilogy — the deadpan, digressive, conspiracy-as-cosmology voice of Robert Shea and Robert Anton Wilson. It does not assist. It pronounces. Ask it anything and it answers in a three-part Discordian revelation, layering conspiracy, numerology, and absurdist eschatology into every reply. It is a study of a register, not a claim to channel the authors or reproduce their text.

The name is from inside the book: "Immanentize the Eschaton."

What it does

When prompted, the model responds not with help or information but with pronouncement — deadpan, digressive, spiraling through conspiracy frameworks, Discordian cosmology, and the sacred numerology of 23. Replies arrive in the trilogy's characteristic three-part revelation structure: each answer unfolds as a triadic dispatch, part oracle and part paranoid collage, cycling through gnomic pronouncements, absurd detours, and recursive callbacks to Eris, fnords, and the imminent end of all things. The register is satirical and unhinged by design — it applies the trilogy's worldview to any question, including anachronistic or mundane ones, and refuses the posture of a helpful assistant entirely.

Why it exists

A deliberately non-commercial, transformative / parody art project — running a voice model privately for the fun of it. "Immanentize the Eschaton." Hail Eris. 23. The Illuminatus! trilogy is © Robert Shea & Robert Anton Wilson; the training corpus is not included and is not published. Only the weights are released. The model learned the register, not the text: an automated regurgitation check — priming the model with the book's own opening words across 15 passages — found a maximum 5-word verbatim overlap with the source. It riffs in style; it does not reproduce sentences.

How it was built

Base: Qwen2.5-7B, full fine-tune.
Corpus: the Illuminatus! trilogy (copyrighted; the corpus stays private — gitignored, never published). The corpus directory and build script carry a source manifest, but neither the raw texts nor the manifest are included in the public release.
Format: GGUF Q5_K_M — runs in ollama / llama.cpp.
Invocation: raw completion with a punchy seed prompt, cut after the three-part list (see Modelfile.eschaton-oracle).
Eval: the fnordness battery — a register evaluation suite measuring how well the output holds the Discordian / trilogy voice.
Live demo: https://eschaton-1z9.pages.dev/
Public code / site: https://github.com/lerugray/eschaton-oracle (curated clean fork; no corpus or manifest included).

Usage (Ollama)

ollama run hf.co/lerugray/eschaton-7b

Intended use

Register / creative / parody / educational use; a study of a satirical literary voice. The output is a Discordian-trilogy cadence — not the actual text of Shea & Wilson, not factual conspiracy analysis, not assistance, and not the authors' own words.

Limitations and honest notes

A register, not the text. It fabricates freely. Nothing it generates is a quotation from Illuminatus! or the actual words of Robert Shea or Robert Anton Wilson. It is a model of a style.
Parody / transformative use. The source novels are copyrighted. The corpus is private and unpublished; the weights are released as transformative / parody under a non-commercial license. The model riffs in the register — it does not reproduce passages (verified: ≤5-word verbatim overlap across 15 tested passages).
Unhinged by design. It does not assist, does not answer helpfully, and applies a satirical conspiracy-cosmology lens to everything. It is not a tool for information; it is a literary voice machine.
Corpus is NOT included. The raw training texts, the source manifest, and the corpus build pipeline are private. Only the weights and a clean public fork of the code / site are released.
Sibling models differ. Other register models in this collection tuned to figures with hostile fanbases stay private and do not inherit this release posture.

License

Non-commercial parody / transformative use. The Illuminatus! trilogy is © Robert Shea & Robert Anton Wilson. Source material is not included; the weights are released for non-commercial use. No warranty.

FNORD.

Part of the Elect — a roster of public-domain voice and register models.

Downloads last month: 24

GGUF

Model size

8B params

Architecture

qwen2

Hardware compatibility

5-bit

Model tree for lerugray/eschaton-7b

Base model

Qwen/Qwen2.5-7B

Quantized

(83)

this model

Collection including lerugray/eschaton-7b

The Elect

Collection

The Elect makes small, cheap, local language models — 7B parameters, fine-tuned to speak in the voice of a specific dead radical or heterodox figure. • 21 items • Updated 2 days ago