Instructions to use sphaela/Qwen3.6-27B-AutoRound-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sphaela/Qwen3.6-27B-AutoRound-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="sphaela/Qwen3.6-27B-AutoRound-GGUF",
	filename="Qwen3.6-27B-Q2_K_MIXED.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use sphaela/Qwen3.6-27B-AutoRound-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf sphaela/Qwen3.6-27B-AutoRound-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 sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf sphaela/Qwen3.6-27B-AutoRound-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 sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M

Use Docker

docker model run hf.co/sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use sphaela/Qwen3.6-27B-AutoRound-GGUF with Ollama:
```
ollama run hf.co/sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M
```

Unsloth Studio new

How to use sphaela/Qwen3.6-27B-AutoRound-GGUF 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 sphaela/Qwen3.6-27B-AutoRound-GGUF 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 sphaela/Qwen3.6-27B-AutoRound-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for sphaela/Qwen3.6-27B-AutoRound-GGUF to start chatting

Pi new

How to use sphaela/Qwen3.6-27B-AutoRound-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M

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": "sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use sphaela/Qwen3.6-27B-AutoRound-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf sphaela/Qwen3.6-27B-AutoRound-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 sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use sphaela/Qwen3.6-27B-AutoRound-GGUF with Docker Model Runner:
```
docker model run hf.co/sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M
```

Lemonade

How to use sphaela/Qwen3.6-27B-AutoRound-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull sphaela/Qwen3.6-27B-AutoRound-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.6-27B-AutoRound-GGUF-Q4_K_M

List all available models

lemonade list

Impressive

by Throghar - opened about 1 month ago

Discussion

Throghar

about 1 month ago

Tested Q2_K mixed, greater speed and so consistent answers compared to tradditional llama.cpp Q4_K_M quants, i would like to see more models in this qunatization method.

sphaela

Owner about 1 month ago

Thanks for the kind words!

tooltd

30 days ago

Why is Qwen3.6-27B-Q2_K_MIXED.gguf larger than Q4_K_M? I've noticed that your other models are usually smaller.
I'm looking for something that's best with 16GB of VRAM :)

sphaela

Owner 30 days ago

Why is Qwen3.6-27B-Q2_K_MIXED.gguf larger than Q4_K_M? I've noticed that your other models are usually smaller.
I'm looking for something that's best with 16GB of VRAM :)

Sorry that it didn't fit in 16GB. I've noticed that dense models will have larger size while MoE models will be smaller. Should be a AutoRound thing.

Throghar

25 days ago

I think it is because more layers are preseved at higher precision because they are marked as important.

BazsiBazsi

20 days ago

Why is Qwen3.6-27B-Q2_K_MIXED.gguf larger than Q4_K_M? I've noticed that your other models are usually smaller.
I'm looking for something that's best with 16GB of VRAM :)

Hi! Depending on your use case, you can fit it to your gpu just reduce batch and ubatch size to 256 in llama.cpp!
This will make a massive hit to your speeds, but i was able to fit Q3_K_L to 12 gm vram with 8k context(q4_1 kv cache) using this method. I had about 30 mb leftover.

meigami

19 days ago

Hi, What's the quantization setting for Q2_K_MIXED? is it better than Q4_K_M?

tooltd

19 days ago

I am running Q4_K_M with 95K context , KV Cache only q4_0 . output is quite stable, feels better than IQ4_XS and UD_Q3_K_XL

sphaela

Owner 18 days ago

•

edited 18 days ago

Hi, What's the quantization setting for Q2_K_MIXED? is it better than Q4_K_M?

Hi, the layers is quantized to a mixed of Q4 and Q2. Overall it should be faster inference than Q4_K_M but with mostly the same quality.

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment