Instructions to use carlosmm26/Atanor-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use carlosmm26/Atanor-4B with HERMES:

# No code snippets available yet for this library.

# To use this model, check the repository files and the library's documentation.

# Want to help? PRs adding snippets are welcome at:
# https://github.com/huggingface/huggingface.js

llama-cpp-python

How to use carlosmm26/Atanor-4B with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="carlosmm26/Atanor-4B",
	filename="atanor-4b-full-Q4_K_M-f16.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 carlosmm26/Atanor-4B 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 carlosmm26/Atanor-4B:F16
# Run inference directly in the terminal:
llama cli -hf carlosmm26/Atanor-4B:F16

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf carlosmm26/Atanor-4B:F16
# Run inference directly in the terminal:
llama cli -hf carlosmm26/Atanor-4B:F16

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 carlosmm26/Atanor-4B:F16
# Run inference directly in the terminal:
./llama-cli -hf carlosmm26/Atanor-4B:F16

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 carlosmm26/Atanor-4B:F16
# Run inference directly in the terminal:
./build/bin/llama-cli -hf carlosmm26/Atanor-4B:F16

Use Docker

docker model run hf.co/carlosmm26/Atanor-4B:F16

LM Studio
Jan

vLLM

How to use carlosmm26/Atanor-4B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "carlosmm26/Atanor-4B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "carlosmm26/Atanor-4B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/carlosmm26/Atanor-4B:F16

Ollama
How to use carlosmm26/Atanor-4B with Ollama:
```
ollama run hf.co/carlosmm26/Atanor-4B:F16
```

Unsloth Studio

How to use carlosmm26/Atanor-4B 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 carlosmm26/Atanor-4B 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 carlosmm26/Atanor-4B to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for carlosmm26/Atanor-4B to start chatting

How to use carlosmm26/Atanor-4B with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf carlosmm26/Atanor-4B:F16

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": "carlosmm26/Atanor-4B:F16"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use carlosmm26/Atanor-4B with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf carlosmm26/Atanor-4B:F16

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 carlosmm26/Atanor-4B:F16

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use carlosmm26/Atanor-4B with Docker Model Runner:
```
docker model run hf.co/carlosmm26/Atanor-4B:F16
```

Lemonade

How to use carlosmm26/Atanor-4B with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull carlosmm26/Atanor-4B:F16

Run and chat with the model

lemonade run user.Atanor-4B-F16

List all available models

lemonade list

🜂 Atanor-4B

A fine-tune of Qwen3.5-4B specialized for agentic tool-use inside Hermes Agent.

There were 9B and 27B versions of Qwen fine-tuned to run as agents in Hermes (the Carnice models). That made me wonder: could a smaller model do the same job?

Atanor-4B is my answer — and my first fine-tune ever. It was trained entirely locally, on a single RTX 3090.

The name Atanor is the alchemist's furnace: the 4B is the lead that goes in, the agent is what comes out. 🜂

Results — agentic evaluation

Measured on a 60-task Hermes-native agent benchmark (real tool execution inside Hermes Agent, deterministic / temperature 0), base vs fine-tune:

Metric	Qwen3.5-4B (base)	Atanor-4B
Agent score	0.81	0.84
Picking the right tool	30%	60% ⬆️ doubled
Task success	67%	73%

The core agent skill — choosing the correct tool for a task — doubled (30% → 60%).

How it was made

Following the Carnice recipe, in two LoRA stages (BF16, on one RTX 3090):

Stage A — reasoning repair: Bespoke-Stratos + NuminaMath-CoT
Stage B — Hermes tool-use: the kai-os/carnice-glm5-hermes-traces traces (the full set, seq len 16384)

~33 hours of training, zero crashes.

Files in this repo

File	What it is
`atanor-4b-full-Q4_K_M.gguf`	Quantized (~2.6 GB) — run it directly in llama.cpp / Hermes / Ollama
`atanor-4b-full-Q4_K_M-f16.gguf`	Full-precision GGUF (~8 GB) — for re-quantizing or lossless inference
`*.safetensors` (merged)	Full merged model for `transformers` / further fine-tuning
`adapter/`	The LoRA adapter alone, to apply on the base model

Usage (llama.cpp / Hermes)

# llama.cpp server
llama-server -m atanor-4b-full-Q4_K_M.gguf --jinja -ngl 99 -c 32768 --alias atanor

# point a Hermes profile at it (provider base_url: http://localhost:8081/v1)
hermes chat --profile atanor -q "read data.csv and total the 'south' region using the terminal"

Thinking is on by default — it helps the model reason about which tool to use. (Pass chat_template_kwargs: {"enable_thinking": false} to disable.)

This is my first fine-tune, and version one. More to come. I learned a ton — and the best part is it was all done at home, on my own GPU.

Downloads last month: 70

Safetensors

Model size

5B params

Tensor type

BF16

F32

Model tree for carlosmm26/Atanor-4B

Base model

Qwen/Qwen3.5-4B-Base

Finetuned

Qwen/Qwen3.5-4B

Adapter

(252)

this model

Adapters

2 models

Datasets used to train carlosmm26/Atanor-4B

Evaluation results

Tool selection on Atanor Hermes Agent eval
self-reported

60.000
Task success on Atanor Hermes Agent eval
self-reported

72.900
Agent score on Atanor Hermes Agent eval
self-reported

0.836