Instructions to use alfonsodlg/celulares-gemma3-4b-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use alfonsodlg/celulares-gemma3-4b-gguf with Transformers:

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("alfonsodlg/celulares-gemma3-4b-gguf", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use alfonsodlg/celulares-gemma3-4b-gguf 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 alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf alfonsodlg/celulares-gemma3-4b-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 alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf alfonsodlg/celulares-gemma3-4b-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 alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M

Use Docker

docker model run hf.co/alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M

LM Studio
Jan
Ollama
How to use alfonsodlg/celulares-gemma3-4b-gguf with Ollama:
```
ollama run hf.co/alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M
```
Unsloth Desktop
Docker Model Runner
How to use alfonsodlg/celulares-gemma3-4b-gguf with Docker Model Runner:
```
docker model run hf.co/alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M
```

Lemonade

How to use alfonsodlg/celulares-gemma3-4b-gguf with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull alfonsodlg/celulares-gemma3-4b-gguf:Q4_K_M

Run and chat with the model

lemonade run user.celulares-gemma3-4b-gguf-Q4_K_M

List all available models

lemonade list

Atomic Chat

celulares-gemma3-4b-gguf / README.md

alfonsodlg

Upload folder using huggingface_hub

6a61b60 verified 9 months ago

preview code

Raw

History Blame Contribute Delete

2.06 kB

metadata

license: apache-2.0
language:
  - es
library_name: transformers
tags:
  - gemma
  - gguf
  - sales
  - conversational
  - spanish
  - fine-tuned
base_model: google/gemma-3-4b-it
model_type: gemma3
quantization: Q4_K_M

🤖 Celulares Gemma 3 4B v3 - Agente de Ventas (GGUF Q4_K_M)

Modelo Gemma 3 4B fine-tuned para ventas de celulares en español peruano.

📊 Especificaciones

Parámetro	Valor
Modelo base	google/gemma-3-4b-it
Cuantización	Q4_K_M (4-bit)
Tamaño	2.4 GB
Idioma	Español (Perú)
Contexto	4,096 tokens
Dataset	1,017 conversaciones

🎯 Uso

Con Ollama

# Descargar el archivo GGUF
wget https://huggingface.co/alfonsodlg/celulares-gemma3-4b-gguf/resolve/main/celulares-v3-q4_k_m.gguf

# Crear Modelfile
cat > Modelfile << 'MODELFILE'
FROM ./celulares-v3-q4_k_m.gguf

TEMPLATE "<start_of_turn>user
{{ .Prompt }}<end_of_turn>
<start_of_turn>model
{{ .Response }}<end_of_turn>
"

SYSTEM "Eres un agente de ventas experto en celulares de Coolbox Perú. Tu personalidad es entusiasta, consultiva y orientada a resultados."

PARAMETER temperature 0.1
PARAMETER top_p 0.9
PARAMETER num_ctx 4096
PARAMETER stop "<end_of_turn>"
MODELFILE

# Importar a Ollama
ollama create celulares:v3-q4 -f Modelfile

Ejemplo de uso

import requests

response = requests.post('http://localhost:11434/api/generate', json={
    'model': 'celulares:v3-q4',
    'prompt': 'Busco un celular Samsung hasta 1500 soles',
    'stream': False
})
print(response.json()['response'])

📈 Métricas

Tasa de éxito: 100% (9/9 tests críticos)
Latencia: 3-8 segundos (respuesta completa)
RAM: ~4GB

🏋️ Entrenamiento

Plataforma: Modal (H100)
Framework: Unsloth
Dataset: 1,017 conversaciones de ventas
Steps: 200
Loss final: ~0.4

📝 Licencia

Apache 2.0

alfonsodlg
/

celulares-gemma3-4b-gguf