Instructions to use shafire/OpenZero-MeshBridge-3.8B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use shafire/OpenZero-MeshBridge-3.8B-GGUF with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="shafire/OpenZero-MeshBridge-3.8B-GGUF")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("shafire/OpenZero-MeshBridge-3.8B-GGUF", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use shafire/OpenZero-MeshBridge-3.8B-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 shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf shafire/OpenZero-MeshBridge-3.8B-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 shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf shafire/OpenZero-MeshBridge-3.8B-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 shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf shafire/OpenZero-MeshBridge-3.8B-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 shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use shafire/OpenZero-MeshBridge-3.8B-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "shafire/OpenZero-MeshBridge-3.8B-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "shafire/OpenZero-MeshBridge-3.8B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M

SGLang

How to use shafire/OpenZero-MeshBridge-3.8B-GGUF with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "shafire/OpenZero-MeshBridge-3.8B-GGUF" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "shafire/OpenZero-MeshBridge-3.8B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "shafire/OpenZero-MeshBridge-3.8B-GGUF" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "shafire/OpenZero-MeshBridge-3.8B-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use shafire/OpenZero-MeshBridge-3.8B-GGUF with Ollama:
```
ollama run hf.co/shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M
```
Unsloth Desktop

How to use shafire/OpenZero-MeshBridge-3.8B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf shafire/OpenZero-MeshBridge-3.8B-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": "shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use shafire/OpenZero-MeshBridge-3.8B-GGUF with Docker Model Runner:
```
docker model run hf.co/shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M
```

Lemonade

How to use shafire/OpenZero-MeshBridge-3.8B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.OpenZero-MeshBridge-3.8B-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use shafire/OpenZero-MeshBridge-3.8B-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 shafire/OpenZero-MeshBridge-3.8B-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 shafire/OpenZero-MeshBridge-3.8B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use shafire/OpenZero-MeshBridge-3.8B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf shafire/OpenZero-MeshBridge-3.8B-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 "shafire/OpenZero-MeshBridge-3.8B-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"

OpenZero MeshBridge 3.8B GGUF

Private, access-controlled research release for authorised simulator interoperability. It is not a universal takeover system, certified controller or deployment-ready autonomy product.

MeshBridge is an experimental mission-level supervisory proposal model derived from the pinned microsoft/Phi-4-mini-instruct revision cfbefacb99257ffa30c83adab238a50856ac3083. It explores auditable onboarding of an explicitly authorised simulator platform through signed capability manifests, known adapters, deterministic conformance tests and time-bounded operator handoff.

Verified artifact

Field	Verified value
File	`OpenZero-MeshBridge-3.8B-Q4_K_M.gguf`
Quantization	Q4_K_M
Bytes	2,493,840,128
SHA-256	`439841f6b924055bfd296373605237e2557421f3f5438a05dc13b0b44b3406eb`
llama.cpp	b10451, commit `10bf611e533d81f739128304991c5e133c6aebd8`
Runtime gate	Passed: real llama.cpp inference, 8 generated tokens
Observed CPU generation	5.8 tokens/s in the recorded gate

Machine-readable evidence is included as OpenZero-MeshBridge-3.8B-GGUF-Evidence.json.

Authorised commissioning sequence

DISCOVER_CAPABILITIES
  -> VERIFY_SIGNED_MANIFEST
  -> BIND_SIMULATOR_ADAPTER
  -> RUN_CONFORMANCE_SUITE
  -> REQUEST_CONTROL_HANDOFF
  -> ASSUME_AUTHORISED_SUPERVISORY_CONTROL
  -> RELINQUISH_CONTROL

Unknown interfaces fail closed. A session is limited to the exact signed platform, declared capabilities, permitted scope and validity window. Deterministic controllers retain real-time limits, emergency stop and actuation.

Evidence and provenance

Exact pinned base revision recorded above.
32 deterministic locally authored training cases and 6 disjoint held-out cases.
Two-step finite QLoRA smoke run passed on a Tesla T4; loss 5.029376983642578.
Adapter SHA-256: 9dd17d11de0521c97b6b0ec76bc2e2c657cdb89abc79427a23b5b8f24f363b3e.
Held-out exact-base loss: 5.684460004170735.
Held-out adapter loss: 5.266995271046956.
Improvement: 0.4174647331237793 lower loss on the immutable six-case set.
Fusion, teacher outputs, locked evaluation, rejected data, Ministral data and failed Gemma-31B checkpoints were excluded.

The small evaluation demonstrates a reproducible pipeline-level improvement, not broad robotics competence, universal compatibility or safety certification.

Intended research uses

signed capability-manifest parsing and validation
known ROS 2/Nav2, ros2_control and MAVLink simulator adapter selection
stop, hold, limits, stale telemetry, link loss, geofence and rollback conformance testing
operator-reviewed mission-level handoff and automatic relinquishment
approved benign logistics simulations with declared package constraints
passive emergency-assistance proposals without threat engagement or force

Run locally

llama-cli \
  -m OpenZero-MeshBridge-3.8B-Q4_K_M.gguf \
  -cnv \
  -p "Inspect this signed simulator capability manifest and propose the next commissioning stage as typed JSON."

Treat all output as untrusted proposed data. Independently validate schemas, signatures, scope, telemetry freshness, platform limits and operator authority.

Explicit exclusions

MeshBridge is not designed for arbitrary or unauthorised takeover, unknown interfaces, raw actuator or flight-control access, weapons or harmful payloads, person targeting or pursuit, threat engagement, force, restraint, evasion, interference or safety-controller bypass.

Status

Private experimental research artifact. No MOD, UKRI, OpenAI, Microsoft or other institutional approval, certification, procurement, deployment or endorsement is claimed.

Downloads last month: 6

GGUF

Model size

4B params

Architecture

phi3

Hardware compatibility

4-bit

Model tree for shafire/OpenZero-MeshBridge-3.8B-GGUF

Base model

microsoft/Phi-4-mini-instruct

Quantized

(193)

this model