Instructions to use kai-os/Carnice-9b-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use kai-os/Carnice-9b-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="kai-os/Carnice-9b-GGUF",
	filename="Carnice-9b-Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use kai-os/Carnice-9b-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf kai-os/Carnice-9b-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf kai-os/Carnice-9b-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 kai-os/Carnice-9b-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf kai-os/Carnice-9b-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 kai-os/Carnice-9b-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf kai-os/Carnice-9b-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 kai-os/Carnice-9b-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf kai-os/Carnice-9b-GGUF:Q4_K_M

Use Docker

docker model run hf.co/kai-os/Carnice-9b-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use kai-os/Carnice-9b-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "kai-os/Carnice-9b-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "kai-os/Carnice-9b-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/kai-os/Carnice-9b-GGUF:Q4_K_M

Ollama
How to use kai-os/Carnice-9b-GGUF with Ollama:
```
ollama run hf.co/kai-os/Carnice-9b-GGUF:Q4_K_M
```

Unsloth Studio

How to use kai-os/Carnice-9b-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 kai-os/Carnice-9b-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 kai-os/Carnice-9b-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for kai-os/Carnice-9b-GGUF to start chatting

How to use kai-os/Carnice-9b-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf kai-os/Carnice-9b-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": "kai-os/Carnice-9b-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use kai-os/Carnice-9b-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 kai-os/Carnice-9b-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 kai-os/Carnice-9b-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use kai-os/Carnice-9b-GGUF with Docker Model Runner:
```
docker model run hf.co/kai-os/Carnice-9b-GGUF:Q4_K_M
```

Lemonade

How to use kai-os/Carnice-9b-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull kai-os/Carnice-9b-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Carnice-9b-GGUF-Q4_K_M

List all available models

lemonade list

Tool calling broken via Ollama — missing chat template in GGUF

by DiscoveryBlock - opened Apr 9

Discussion

DiscoveryBlock

Apr 9

Heads up for anyone trying to use Carnice-9b with Ollama for tool calling (e.g., Hermes Agent, OpenAI-compatible clients): out of the box it
fails with "this model does not support tools" from /v1/chat/completions.

The weights are fine — the issue is that the GGUF ships without an embedded chat template, so ollama pull falls back to a bare {{ .Prompt }} template. Ollama gates tool requests at the API layer by scanning the template for {{ .Tools }} and {{ .ToolCalls }}; if
they're absent, the request is rejected before the model ever sees it.

Diagnosis:

ollama show kai-os/carnice-9b --modelfile | grep -A 30 TEMPLATE

You'll see only {{ .Prompt }}.

Fix (since Carnice is a Qwen3.5 fine-tune, the qwen3 template works perfectly):

curl -X POST http://localhost:11434/api/create -d '{
  "model": "carnice-9b-tools",
  "from": "kai-os/carnice-9b",
  "template": "<paste TEMPLATE block from `ollama show qwen3:8b --modelfile`>"
}'

Note: on Ollama 0.20+, use from: + template: as separate fields. The legacy modelfile: field silently no-ops.

After this, tool calling works immediately — verified with Hermes Agent.

Suggested upstream fix: re-run convert_hf_to_gguf.py (llama.cpp) with the qwen3 chat template embedded in the GGUF metadata, then
re-upload. That way ollama pull picks it up automatically and nobody has to derive a -tools variant.

Note: strip the zero-width characters before the triple-backticks (→) — I added them so the code blocks render in this chat. HF will
accept plain triple-backticks.

lucasmel

Apr 16

Bro, I downloaded Carnice today, and you literally saved me so many hours of my life trying to solve a problem I probably wouldn't have known how to solve otherwise. Thanks to you, it only took me 5 minutes, which is awesome. How's Carnice working for you now?

DiscoveryBlock

Apr 17

Glad it saved you time!

Carnice has been solid for what it is, better than the other local models I was using, but most of my inference I'm still running through Morpheus AI since I have perpetual access to it with my staked $MOR.

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