Instructions to use Mattimax/DAC6-0.5B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Mattimax/DAC6-0.5B with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Mattimax/DAC6-0.5B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Mattimax/DAC6-0.5B")
model = AutoModelForCausalLM.from_pretrained("Mattimax/DAC6-0.5B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Mattimax/DAC6-0.5B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Mattimax/DAC6-0.5B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Mattimax/DAC6-0.5B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Mattimax/DAC6-0.5B

SGLang

How to use Mattimax/DAC6-0.5B with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Mattimax/DAC6-0.5B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Mattimax/DAC6-0.5B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Mattimax/DAC6-0.5B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Mattimax/DAC6-0.5B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Mattimax/DAC6-0.5B with Docker Model Runner:
```
docker model run hf.co/Mattimax/DAC6-0.5B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

☕ Support the project

Model Card: DAC5‑0.5B (M.INC. Research)

DAC6 Logo

DAC5‑0.6B (M.INC. Research)

🇮🇹 Model Card – DAC6

📘 Panoramica

DAC6 è un modello linguistico di 500 milioni di parametri (0.5B), progettato per offrire un equilibrio tra efficienza e capacità di ragionamento.
È sviluppato principalmente per l’ecosistema Android, dove può conversare fluentemente e eseguire chiamate a strumenti locali (tool calling) direttamente sul dispositivo.

🧠 Caratteristiche principali

Lingua principale: Italiano 🇮🇹
Lingua secondaria: Inglese 🇬🇧
Capacità: Conversazione naturale, risposte emotive, tool calling locale
Ambiente target: Android (locale, sperimentale)
Stato: In fase di sviluppo — versione sperimentale, soggetta a instabilità e ottimizzazioni continue
Dimensione: 500M parametri (0.5B)
Obiettivo: Fornire un’esperienza fluida e naturale, con capacità di interazione diretta con funzioni del dispositivo

⚙️ Note tecniche

Architettura ottimizzata per dispositivi mobili
Supporto a tool calling per funzioni locali (es. sensori, app, moduli AI)
Addestrato su dataset misti con prevalenza di lingua italiana e contenuti multidominio

🇬🇧 Model Card – DAC6

📘 Overview

DAC6 is a 500‑million‑parameter (0.5B) language model designed for balanced performance and reasoning efficiency.
It’s primarily developed for the Android ecosystem, capable of natural conversation and local tool calling directly on the device.

🧠 Key Features

Primary language: Italian 🇮🇹
Secondary language: English 🇬🇧
Capabilities: Conversational fluency, emotional responses, local tool calling
Target environment: Android (local, experimental)
Status: Under development — experimental version, with some instabilities expected
Size: 500M parameters (0.5B)
Goal: Deliver smooth, natural interaction and direct integration with device functions

⚙️ Technical Notes

Architecture optimized for mobile devices
Supports local tool calling (e.g., sensors, apps, AI modules)
Trained on mixed datasets with a strong Italian language focus and multidomain content

Downloads last month: 31

Safetensors

Model size

0.5B params

Tensor type

BF16

Model tree for Mattimax/DAC6-0.5B

Quantizations

1 model

Collection including Mattimax/DAC6-0.5B

DAC6

Collection

A complete collection with all DAC6 Models. • 1 item • Updated 4 days ago