Instructions to use ubermorgen/qwen3-devops with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use ubermorgen/qwen3-devops 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 ubermorgen/qwen3-devops
# Run inference directly in the terminal:
llama cli -hf ubermorgen/qwen3-devops

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf ubermorgen/qwen3-devops
# Run inference directly in the terminal:
llama cli -hf ubermorgen/qwen3-devops

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 ubermorgen/qwen3-devops
# Run inference directly in the terminal:
./llama-cli -hf ubermorgen/qwen3-devops

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 ubermorgen/qwen3-devops
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ubermorgen/qwen3-devops

Use Docker

docker model run hf.co/ubermorgen/qwen3-devops

LM Studio
Jan

vLLM

How to use ubermorgen/qwen3-devops with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ubermorgen/qwen3-devops"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ubermorgen/qwen3-devops",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/ubermorgen/qwen3-devops

Ollama
How to use ubermorgen/qwen3-devops with Ollama:
```
ollama run hf.co/ubermorgen/qwen3-devops
```

Unsloth Studio

How to use ubermorgen/qwen3-devops 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 ubermorgen/qwen3-devops 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 ubermorgen/qwen3-devops to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for ubermorgen/qwen3-devops to start chatting

How to use ubermorgen/qwen3-devops with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ubermorgen/qwen3-devops

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "ubermorgen/qwen3-devops"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

OpenClaw new

How to use ubermorgen/qwen3-devops with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ubermorgen/qwen3-devops

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 "ubermorgen/qwen3-devops" \
  --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"

Docker Model Runner
How to use ubermorgen/qwen3-devops with Docker Model Runner:
```
docker model run hf.co/ubermorgen/qwen3-devops
```

Lemonade

How to use ubermorgen/qwen3-devops with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ubermorgen/qwen3-devops

Run and chat with the model

lemonade run user.qwen3-devops-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use ubermorgen/qwen3-devops with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ubermorgen/qwen3-devops

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 ubermorgen/qwen3-devops

Run Hermes

hermes

Atomic Chat

Ollama DevOps Agent

A lightweight AI-powered DevOps automation tool using a fine-tuned Qwen3-1.7B model with Ollama and SmolAgents. Specialized for Docker and Kubernetes workflows with sequential tool execution and structured reasoning.

Features

Sequential Tool Execution: Calls ONE tool at a time, waits for results, then proceeds
Structured Reasoning: Uses <think> and <plan> tags to show thought process
Validation-Aware: Checks command outputs for errors before proceeding
Multi-Step Tasks: Handles complex workflows requiring multiple tool calls
Approval Mode: User confirmation before executing each tool call for enhanced safety (enabled by default)
Resource Efficient: Optimized for local development (1GB GGUF model)
Fast: Completes typical DevOps tasks in ~10 seconds

What's Special About This Model?

This model is fine-tuned specifically for DevOps automation with improved reasoning capabilities:

Docker & Kubernetes Expert: Trained on 300+ Docker and Kubernetes workflows (90% of training data)
One tool at a time: Unlike base models that try to call all tools at once, this model executes sequentially
Explicit planning: Shows reasoning with <think> and <plan> before acting
Uses actual values: Extracts and uses real values from tool responses in subsequent calls
Error handling: Validates each step and tries alternative approaches on failure

Training Data Focus

The model has been trained on:

Docker workflows: Building images, containers, Docker Compose, optimization
Kubernetes operations: Pods, deployments, services, configurations
General DevOps: File operations, system commands, basic troubleshooting

⚠️ Note: The model has limited training on cloud-specific CLIs (gcloud, AWS CLI, Azure CLI). For best results, use it for Docker and Kubernetes tasks.

Example Output

Task: Get all pods in default namespace

Step 1: Execute kubectl command
<tool_call>
{"name": "bash", "arguments": {"command": "kubectl get pods -n default"}}
</tool_call>

[Receives pod list]

Step 2: Provide summary
<tool_call>
{"name": "final_answer", "arguments": {"answer": "Successfully retrieved 10 pods in default namespace..."}}
</tool_call>

Quick Start

🎯 Recommended: Native Installation

For the best experience with full DevOps capabilities:

curl -fsSL https://raw.githubusercontent.com/ubermorgenland/devops-agent/main/install.sh | bash

This will automatically:

Install Ollama (if not present)
Install Python dependencies
Download the model from Hugging Face
Create the Ollama model
Set up the devops-agent CLI command

Why native installation?

✅ Full system access - manage real infrastructure
✅ No credential mounting - works with your existing setup
✅ Better performance - no container overhead
✅ Simpler usage - just run devops-agent

Downloads last month: 44

GGUF

Model size

2B params

Architecture

qwen3

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Model tree for ubermorgen/qwen3-devops

Base model

Qwen/Qwen3-1.7B-Base

Finetuned

Qwen/Qwen3-1.7B

Quantized

(326)

this model