Instructions to use espetro/kev-9b-demo-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 espetro/kev-9b-demo-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 espetro/kev-9b-demo-gguf
# Run inference directly in the terminal:
llama cli -hf espetro/kev-9b-demo-gguf

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf espetro/kev-9b-demo-gguf
# Run inference directly in the terminal:
llama cli -hf espetro/kev-9b-demo-gguf

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 espetro/kev-9b-demo-gguf
# Run inference directly in the terminal:
./llama-cli -hf espetro/kev-9b-demo-gguf

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 espetro/kev-9b-demo-gguf
# Run inference directly in the terminal:
./build/bin/llama-cli -hf espetro/kev-9b-demo-gguf

Use Docker

docker model run hf.co/espetro/kev-9b-demo-gguf

LM Studio
Jan
Ollama
How to use espetro/kev-9b-demo-gguf with Ollama:
```
ollama run hf.co/espetro/kev-9b-demo-gguf
```
Unsloth Desktop

How to use espetro/kev-9b-demo-gguf with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf espetro/kev-9b-demo-gguf

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": "espetro/kev-9b-demo-gguf"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use espetro/kev-9b-demo-gguf with Docker Model Runner:
```
docker model run hf.co/espetro/kev-9b-demo-gguf
```

Lemonade

How to use espetro/kev-9b-demo-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull espetro/kev-9b-demo-gguf

Run and chat with the model

lemonade run user.kev-9b-demo-gguf-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use espetro/kev-9b-demo-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 espetro/kev-9b-demo-gguf

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 espetro/kev-9b-demo-gguf

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use espetro/kev-9b-demo-gguf with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf espetro/kev-9b-demo-gguf

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 "espetro/kev-9b-demo-gguf" \
  --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"

Kev-9B v1.0 (demo quant) - GGUF

Packed Kev "System One" decision model: Qwen3.5-9B backbone + trained pointer head, quantized for fast downloads. Answers typed questions (noul, choice, score) about a state with calibrated probabilities - no text generation.

This is a demo quantization. ~5.6 GB download for quick trials. Probabilities can drift vs the full-precision model and confidence is approximate - it shows how System One answers work, not what a production deployment should ship. For the best-calibrated artifact use espetro/kev-9b-gguf (q8_0 + f16).

Built with the Kev fork of llama.cpp: q4_k_m + a task-specific importance matrix (the token_embd override is already the default at 9B). Measured vs packed F16 on a 17-question probe set: 0 argmax flips, mean |dp| 0.023, max |dp| 0.136.

Use it

# server - TypeSafe /v1/systemone API
llama-server -hf espetro/kev-9b-demo-gguf

# CLI
llama-decide -hf espetro/kev-9b-demo-gguf --json request.json

Or try it in your browser right now - it runs fully client-side via WebAssembly:

https://espetro.github.io/llama.cpp/?model=https://huggingface.co/espetro/kev-9b-demo-gguf/resolve/main/kev-9b-demo-q4km-im.gguf

Source

Weights: jaredpalmer/kev-9b v1.0 adapter merged onto Qwen/Qwen3.5-9B-Base, packed with tools/kev/kev_pack.py, then quantized (llama-imatrix on a decision-input corpus, llama-quantize --imatrix ... q4_k_m). Kev/Jev research: https://github.com/jaredpalmer/kev

Downloads last month: 284

GGUF

Model size

9B params

Architecture

qwen35

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Inference Providers NEW

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