Instructions to use alfonsodlg/celulares-gemma3-4b-gguf with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
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
| 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 | |
| ```bash | |
| # 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 | |
| ```python | |
| 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 | |
| ## 🔗 Enlaces | |
| - [Repositorio del proyecto](https://github.com/alfonsodlg/pricex-smart) | |
| - [Modelo base Gemma 3](https://huggingface.co/google/gemma-3-4b-it) | |