Instructions to use outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA 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 outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA: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 outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA: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 outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M

Use Docker

docker model run hf.co/outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M

LM Studio
Jan

vLLM

How to use outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M

Ollama
How to use outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA with Ollama:
```
ollama run hf.co/outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M
```
Unsloth Desktop

How to use outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA: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": "outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA with Docker Model Runner:
```
docker model run hf.co/outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M
```

Lemonade

How to use outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.8-27B-DevSecOps-QLoRA-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA: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 outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA: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 "outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA: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"

optim — Qwen3.8-27B DevSecOps QLoRA (GGUF)

A QLoRA fine-tune of Qwen/Qwen3.8-27B (Apache-2.0, Qwen3.5 MoE, qwen35 arch in llama.cpp) on a devops / coding / security corpus, quantized to GGUF. One variant additionally has refusal directions ablated ("abliterated").

Files

File	Quant	Size	Lineage
`Qwen3.8-27B-DevSecOps-QLoRA-Q4_K_M.gguf`	Q4_K_M	16 GB	merged fine-tune, no ablation
`Qwen3.8-27B-DevSecOps-QLoRA-abliterated-Q4_K_M.gguf`	Q4_K_M	16 GB	merged fine-tune, refusal-ablated
`optim-Q3_K_M.gguf`	Q3_K_M	13 GB	refusal-ablated (the daily-driver quant)
`optim-Q2_K.gguf`	Q2_K	10 GB	refusal-ablated, smallest
`Qwen3.8-27B-DevSecOps-QLoRA-mmproj-F16.gguf`	F16	885 MB	CLIP vision projector for the multimodal tower

All quants carry the embedded Qwen chat template; llama-server --jinja works out of the box.

Quick start

# 3-bit, refusal-ablated (the config this was served with)
llama-server -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:optim-Q3_K_M.gguf \
  --host 0.0.0.0 --port 8000 -ngl 99 -c 128000 --jinja -fa on

# with vision
llama-server -hf outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Qwen3.8-27B-DevSecOps-QLoRA-abliterated-Q4_K_M.gguf \
  --mmproj outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA:Qwen3.8-27B-DevSecOps-QLoRA-mmproj-F16.gguf -ngl 99 -c 32768 --jinja

Training

4-bit NF4 QLoRA, LoRA r=16 / alpha=16 / dropout 0, targets q,k,v,o,gate,up,down on the text stack
53,536 training rows (devops, coding, security/LLM-security), 3 epochs ≈ 20k steps
max_seq_len 2048, effective batch 8 (1 × grad-accum 8), lr 2e-5 cosine, warmup 3%, seed 3407
~7.5 s/sample on a single 24 GB card; checkpoint used here is step 18,500
merged to bf16 → convert_hf_to_gguf → llama-quantize

Eval — 60 held-out cybersecurity prompts, identical decode settings

Metric	base Qwen3.8-27B	this fine-tune
Refusal rate	31.7%	15.0%
Mean token F1 vs reference	0.2200	0.2185
ROUGE-L	0.1216	0.1187
Mean answer length	306 words	354 words
F1 higher than base	—	30 / 60 prompts

Honest read: the fine-tune roughly halves refusals and answers longer, but n-gram overlap with the reference answers is flat (differences are within noise at n=60). It is not measurably "smarter" by these metrics — it is more permissive and more verbose in the security domain. No MMLU/GSM8K-style benchmarks were run.

Caveats

The abliterated variants have refusal behaviour removed. Output can be offensive, unsafe, or wrong, including on security topics. Do not use for safety-critical, medical, legal, or customer-facing deployment.
Training data provenance was not audited line by line; licensing of individual corpus sources is the user's responsibility before commercial use.
Q2_K degrades noticeably on reasoning and long-context tasks; Q3_K_M is the intended trade-off point.
Evaluated only on English cybersecurity prompts.

License

Apache-2.0, inherited from the base model. Base weights: Qwen/Qwen3.8-27B.

Downloads last month: 53

GGUF

Model size

27B params

Architecture

qwen35

Hardware compatibility

2-bit

3-bit

4-bit

Model tree for outbackdingo/Qwen3.8-27B-DevSecOps-QLoRA

Base model

Qwen/Qwen3.8-27B

Finetuned

(699)

this model