Instructions to use CromIA/CROM-IA-V1-DNA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use CromIA/CROM-IA-V1-DNA with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="CromIA/CROM-IA-V1-DNA",
	filename="crom-dna-1x3-fixo.gguf",
)

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

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use CromIA/CROM-IA-V1-DNA with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf CromIA/CROM-IA-V1-DNA
# Run inference directly in the terminal:
llama-cli -hf CromIA/CROM-IA-V1-DNA

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf CromIA/CROM-IA-V1-DNA
# Run inference directly in the terminal:
llama-cli -hf CromIA/CROM-IA-V1-DNA

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 CromIA/CROM-IA-V1-DNA
# Run inference directly in the terminal:
./llama-cli -hf CromIA/CROM-IA-V1-DNA

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 CromIA/CROM-IA-V1-DNA
# Run inference directly in the terminal:
./build/bin/llama-cli -hf CromIA/CROM-IA-V1-DNA

Use Docker

docker model run hf.co/CromIA/CROM-IA-V1-DNA

LM Studio
Jan

vLLM

How to use CromIA/CROM-IA-V1-DNA with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "CromIA/CROM-IA-V1-DNA"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CromIA/CROM-IA-V1-DNA",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/CromIA/CROM-IA-V1-DNA

Ollama
How to use CromIA/CROM-IA-V1-DNA with Ollama:
```
ollama run hf.co/CromIA/CROM-IA-V1-DNA
```

Unsloth Studio new

How to use CromIA/CROM-IA-V1-DNA 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 CromIA/CROM-IA-V1-DNA 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 CromIA/CROM-IA-V1-DNA to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for CromIA/CROM-IA-V1-DNA to start chatting

Pi new

How to use CromIA/CROM-IA-V1-DNA with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf CromIA/CROM-IA-V1-DNA

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": "CromIA/CROM-IA-V1-DNA"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use CromIA/CROM-IA-V1-DNA with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf CromIA/CROM-IA-V1-DNA

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 CromIA/CROM-IA-V1-DNA

Run Hermes

hermes

Docker Model Runner
How to use CromIA/CROM-IA-V1-DNA with Docker Model Runner:
```
docker model run hf.co/CromIA/CROM-IA-V1-DNA
```

Lemonade

How to use CromIA/CROM-IA-V1-DNA with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull CromIA/CROM-IA-V1-DNA

Run and chat with the model

lemonade run user.CROM-IA-V1-DNA-{{QUANT_TAG}}

List all available models

lemonade list

CROM-IA V1 DNA 🧬

A primeira IA Sub-Simbólica Termodinâmica para Edge Devices

Descrição

Este modelo é um Fine-Tune LoRA do Qwen 2.5 0.5B Instruct, treinado com um dataset especial que codifica instruções em sequências Radix-4 (DNA: A, T, C, G).

O objetivo é permitir inferência de alta performance em hardware extremamente limitado (CPU-only, sem GPU) através de compressão termodinâmica e montagem FUSE com mmap zero-copy.

Métricas de Performance

Métrica	Valor
Hardware	Intel i5-3320M (Ivy Bridge, 2012)
RAM Utilizada	< 700 MB via FUSE mmap
GPU Necessária	Nenhuma (0 layers offloaded)
Velocidade (Prompt)	~37.5 t/s
Velocidade (Geração)	~11.4 t/s
Quantização	Q4_K_M (GGUF)
Tamanho	~380 MB

Como Usar

Opção 1: Via CROM-IA Engine (Recomendado)

git clone https://github.com/MrJc01/crompressor-ia.git
cd crompressor-ia
# Coloque o .gguf na pasta models/
./iniciar_chat_real.sh

Opção 2: Via llama.cpp Direto

./llama-cli \
    -m qwen2.5-crom-dna.gguf \
    --threads 4 \
    -c 512 -b 256 -n 512 \
    --temp 0.1 \
    --repeat_penalty 1.15 \
    -cnv \
    -p "You are CROM-IA, an AI assistant. Respond in Portuguese."

Arquitetura de Treinamento

Dataset: Alpaca-GPT4 PT com filtragem de Entropia de Shannon (H > 7.5 descartado)
Método: LoRA (r=16, alpha=32) sobre Qwen 2.5 0.5B
Codificação DNA: Bytes UTF-8 → pares de 2 bits → bases nitrogenadas (A=00, T=01, C=10, G=11)
Bypass VRAM: Modelo servido via FUSE VFS + mmap, RSS constante

Limitações

Modelo compacto (0.5B params) — ideal para tarefas concisas e diretas
Respostas longas podem perder coerência após ~200 tokens
Otimizado para português brasileiro

Licença

MIT — Livre para uso pessoal, acadêmico e comercial.

Model tree for CromIA/CROM-IA-V1-DNA

Base model

Qwen/Qwen2.5-0.5B

Finetuned

Qwen/Qwen2.5-0.5B-Instruct

Quantized

(207)

this model

CromIA
/

CROM-IA-V1-DNA