Instructions to use sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF",
	filename="Qwen3.6-35B-A3B-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 Settings

llama.cpp

How to use sphaela/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf sphaela/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf sphaela/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf sphaela/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M

Use Docker

docker model run hf.co/sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF with Ollama:
```
ollama run hf.co/sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M
```

Unsloth Studio

How to use sphaela/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-AutoRound-GGUF to start chatting

How to use sphaela/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use sphaela/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-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/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF with Docker Model Runner:
```
docker model run hf.co/sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M
```

Lemonade

How to use sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull sphaela/Qwen3.6-35B-A3B-AutoRound-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Qwen3.6-35B-A3B-AutoRound-GGUF-Q4_K_M

List all available models

lemonade list

Coherence issue

by jesterb32 - opened 5 days ago

Discussion

jesterb32

5 days ago

First off, thanks for these!

Just pulled Q6_K today and ran it with latest llama.cpp release. It seems to be generating repeated \ and that's it. I confirmed that llama.cpp is working. Anyone else seeing this?

Yungchen-Jen

5 days ago

be generating repeated \ and that's it

Same for me. I used Q4_K_M instead.

tooltd

5 days ago

This comment has been hidden (marked as Resolved)

sphaela

Owner 5 days ago

•

edited 5 days ago

Hi thank for reporting, I'm invertigating
Edit: After running all the models under llama.cpp, here's a quantization bug on scheme that's sym=False, data_type=int_asym_dq with super_bits >= 6. So 3 quants (Q4_K_M, Q5_K_M, Q6_K) need to be re-quantized. I'd update the repo shortly. Sorry for the inconvenience.

sphaela

Owner 4 days ago

This seems like a upstream issue of how AutoRound react with MTP, I'm filing a issue to them now. Meanwhile please use other unaffected scheme instead.

soyalemujica

4 days ago

This seems like a upstream issue of how AutoRound react with MTP, I'm filing a issue to them now. Meanwhile please use other unaffected scheme instead.

Does this issue occur with the 27B AutoRound models as well ?

sphaela

Owner 4 days ago

This seems like a upstream issue of how AutoRound react with MTP, I'm filing a issue to them now. Meanwhile please use other unaffected scheme instead.

Does this issue occur with the 27B AutoRound models as well ?

This issue only affect MoE model so the dense 27B is not affected!

sphaela

Owner 4 days ago

Upstream is fixed with #1908 and I'm currently rerunning the quantization. Hold tight!

Yungchen-Jen

4 days ago

@sphaela Thank mate, the fix works now!

sphaela

Owner 4 days ago

This is fixed now :3

soyalemujica

3 days ago

Think you can make a GGUF format of this https://huggingface.co/webhie/Qwen3.6-27B-int4-AutoRound-Code ?

sphaela

Owner 3 days ago

Think you can make a GGUF format of this https://huggingface.co/webhie/Qwen3.6-27B-int4-AutoRound-Code ?

I could do that, but basically I'll need to fine tune from scratch. Since working on their model means dequant back to F16 first and then requant back to GGUF, that will create a lot of loss. If I do it I would start with a better dataset, maybe with the Opus dataset? Or you just want a code specific variant?

jesterb32

3 days ago

Thank you @sphaela !

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