Instructions to use MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF",
	filename="MiniCPM5-1B-lost-frequency-radio-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 MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-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 MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-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 MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-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 MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M

Use Docker

docker model run hf.co/MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-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": "MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M

Ollama
How to use MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF with Ollama:
```
ollama run hf.co/MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M
```

Unsloth Studio

How to use MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-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 MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-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 MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF to start chatting

How to use MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-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": "MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-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 MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-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 MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat new
Docker Model Runner
How to use MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF with Docker Model Runner:
```
docker model run hf.co/MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M
```

Lemonade

How to use MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.MiniCPM5-1B-lost-frequency-radio-GGUF-Q4_K_M

List all available models

lemonade list

MiniCPM5-1B — Lost Frequency Radio (GGUF Q4_K_M)

Fine-tune LoRA de openbmb/MiniCPM5-1B para Lost Frequency Radio, una radio interactiva de universos paralelos construida para el Hugging Face Build Small Hackathon 2026 (track 🍄 An Adventure in Thousand Token Wood).

Demo: Space Lost Frequency Radio
Dataset: ~786 transmisiones de radio surrealistas (es / en / fr) con tokens estructurados, generadas por plantillas y curadas a mano.
Tarea: escribir guiones de radio cortos (60-90 palabras) en personaje: locutores de los años 50, partes meteorológicos de Júpiter, comerciales imposibles, number stations, programas nocturnos entre universos.
Anti-fuga de prompt: los system prompts NO contienen reglas en forma de instrucción ("escribe solo el guion, 60-90 palabras…"); el formato se aprende solo de las completions, así que un modelo de 1B no tiene nada que "recitar" al aire. El francés se enseña a nivel de pesos con su propia tajada del dataset.

Tokens estructurados

El modelo emite marcadores que el frontend convierte en eventos audiovisuales:

Token	Efecto en la radio
`[JINGLE]`	pulso de luz + arpegio
`[INTERFERENCIA]`	glitch de pantalla + ráfaga de estática
`[CORTE COMERCIAL]`	clic + atenuación
`[FIN DE TRANSMISION]`	fade del display y caída de señal

Uso con llama.cpp

from llama_cpp import Llama

llm = Llama(model_path="MiniCPM5-1B-lost-frequency-radio-Q4_K_M.gguf", n_ctx=2048)
prompt = (
    "<s><|im_start|>system\nEres la voz oficial del Servicio Meteorológico de "
    "Júpiter, año 2187. Escribes guiones de radio en español.<|im_end|>\n"
    "<|im_start|>user\nEscribe la transmisión de esta noche. Solo el guion al "
    "aire.<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\n"
)
tokens = llm.tokenize(prompt.encode(), add_bos=False, special=True)
out = llm.create_completion(prompt=tokens, max_tokens=220, temperature=0.7,
                            stop=["<|im_end|>"])
print(out["choices"][0]["text"])

Nota: el prefill <think>\n\n</think>\n\n desactiva el modo razonamiento de MiniCPM5 (equivale a enable_thinking=False del chat template).

Entrenamiento

786 ejemplos (es / en / fr), LoRA r=16, alpha=32, dropout 0.05, sobre todas las proyecciones (q/k/v/o/gate/up/down)
3 épocas, lr 1e-4 cosine, bf16, max_length 768
Hardware: una sola RTX 4050 laptop (6 GB) — el modelo es diminuto a propósito
Loss final ≈ 0.36-0.42, token accuracy ≈ 0.92

Archivos

MiniCPM5-1B-lost-frequency-radio-Q4_K_M.gguf — cuantización Q4_K_M (~651 MB), la que usa el Space
lora-adapter/ — adaptadores LoRA (para reproducir o seguir entrenando)

Downloads last month: -

GGUF

Model size

1B params

Architecture

llama

Hardware compatibility

4-bit

Model tree for MarianaCodebase/MiniCPM5-1B-lost-frequency-radio-GGUF

Base model

openbmb/MiniCPM5-1B

Adapter

(18)

this model

MarianaCodebase
/

MiniCPM5-1B-lost-frequency-radio-GGUF