README.md

---
title: Gemma API
emoji: 🚀
colorFrom: blue
colorTo: red
sdk: docker
app_port: 7860
---

# Gemma 4 E4B OpenAI-Compatible API

This Space runs `google_gemma-4-E4B-it` (Q4_K_M GGUF) via `llama-cpp-python[server]`, providing an OpenAI-compatible API endpoint.

## Endpoint

You can use this just like the OpenAI API by pointing your base URL to the Space URL + `/v1`.

```python
from openai import OpenAI

client = OpenAI(
    base_url="https://cloudunity-gemma-api.hf.space/v1",
    api_key="your_huggingface_token" # Usually ignored by llama.cpp but required by client
)

response = client.chat.completions.create(
    model="gemma-4-E4B-it-Q4_K_M", # Model name is ignored if there's only one, but good practice
    messages=[
        {"role": "user", "content": "Write a short poem about coding."}
    ]
)

print(response.choices[0].message.content)
```