Instructions to use Thox-ai/ThoxUSB-Edge with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use Thox-ai/ThoxUSB-Edge 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 Thox-ai/ThoxUSB-Edge:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Thox-ai/ThoxUSB-Edge:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Thox-ai/ThoxUSB-Edge:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Thox-ai/ThoxUSB-Edge:Q4_K_M

Build from source code

git clone https://github.com/ggml-org/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-app
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama serve -hf Thox-ai/ThoxUSB-Edge:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama cli -hf Thox-ai/ThoxUSB-Edge:Q4_K_M

Use Docker

docker model run hf.co/Thox-ai/ThoxUSB-Edge:Q4_K_M

Llama new
LM Studio
Jan

vLLM

How to use Thox-ai/ThoxUSB-Edge with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/Thox-ai/ThoxUSB-Edge:Q4_K_M

Ollama
How to use Thox-ai/ThoxUSB-Edge with Ollama:
```
ollama run hf.co/Thox-ai/ThoxUSB-Edge:Q4_K_M
```
Unsloth Desktop

How to use Thox-ai/ThoxUSB-Edge with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Thox-ai/ThoxUSB-Edge: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": "Thox-ai/ThoxUSB-Edge:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use Thox-ai/ThoxUSB-Edge with Docker Model Runner:
```
docker model run hf.co/Thox-ai/ThoxUSB-Edge:Q4_K_M
```

Lemonade

How to use Thox-ai/ThoxUSB-Edge with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Thox-ai/ThoxUSB-Edge:Q4_K_M

Run and chat with the model

lemonade run user.ThoxUSB-Edge-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use Thox-ai/ThoxUSB-Edge with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Thox-ai/ThoxUSB-Edge: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 Thox-ai/ThoxUSB-Edge:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use Thox-ai/ThoxUSB-Edge with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf Thox-ai/ThoxUSB-Edge: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 "Thox-ai/ThoxUSB-Edge: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"

ThoxUSB-Edge — On-Device Assistant for Small Edge Devices

Your AI. Your Data. Your Rules.™

Summary

ThoxUSB-Edge is a THOX on-device model (873M parameters) fine-tuned on THOX chat and agent/tool-use corpora for the THOX device program, published as safetensors plus a Q4_K_M GGUF (529 MB) for small edge devices. Status: experimental. No training log, measured benchmark, or on-device validation is published for this build yet.

Files

File	Size	Purpose
`model.safetensors`	1,504,827,608 B	Model weights (safetensors)
`ThoxUSB-Edge.Q4_K_M.gguf`	529,288,992 B	GGUF model build
`tokenizer.json`	19,989,325 B	Tokenizer files
`chat_template.jinja`	7,755 B	Chat template
`config.json`	1,792 B	Model configuration
`.gitattributes`	1,631 B	Repository file
`tokenizer_config.json`	1,125 B	Tokenizer files
`Modelfile`	238 B	Ollama model definition
`generation_config.json`	164 B	Generation configuration

Run it

hf download Thox-ai/ThoxUSB-Edge ThoxUSB-Edge.Q4_K_M.gguf --local-dir ./ThoxUSB-Edge
llama-cli -m ./ThoxUSB-Edge/ThoxUSB-Edge.Q4_K_M.gguf -p "Hello" -n 128

ollama run hf.co/Thox-ai/ThoxUSB-Edge:Q4_K_M

Requires a llama.cpp / Ollama build with support for this architecture. Sized for hosts with at least 1 GB RAM.

Intended use

On-device assistant and tool-routing for the THOX device program.
Out of scope: safety-critical, medical, or legal decisions without a human in the loop; unattended control of physical devices without a confirmation step.

Training data and procedure

THOX fine-tune of Qwen/Qwen3.5-0.8B. The corpus and procedure are described in this card as published at release (see Summary and Evaluation); no training beyond what this card describes is claimed for this repository.

Evaluation

No THOX GATE49 result or benchmark is published for this model. Treat it as experimental.

Limitations

Unvalidated build: behaviour on device has not been measured; verify outputs before relying on them.
Small models make factual mistakes and can produce confident but wrong output.
Behaviour and bias reflect the training data, which is mostly English; quality in other languages is lower.
Quantization trades accuracy for size.

Provenance

Published by THOX.ai as Thox-ai/ThoxUSB-Edge. Upstream base: Qwen/Qwen3.5-0.8B. This card was conformed to THOX Model Card Standard v1.0 on 2026-10-31 (frontmatter, tags, header, files, and section structure); the technical content is as published at release.

License

Apache-2.0. The THOX fine-tune and additions in this repo are covered under Apache-2.0; the Apache-2.0 licence and NOTICE carry forward. See the license text for terms.

THOX.ai — Your AI. Your Data. Your Rules.™

Downloads last month: 55

Safetensors

Model size

0.8B params

Tensor type

BF16

Model tree for Thox-ai/ThoxUSB-Edge

Base model

Qwen/Qwen3.5-0.8B-Base

Finetuned

Qwen/Qwen3.5-0.8B

Finetuned

(605)

this model