app.py

from gradio import Interface
from llama_cpp import Llama

# Load your model
model_path = "models/blobs/sha256-2bada8a7450677000f678be90653b85d364de7db25eb5ea54136ada5f3933730"  # Upload your model file to the Space
llm = Llama(model_path=model_path, n_ctx=4096, verbose=False)

def generate(system_prompt, messages):
    """Generate response from the model"""
    full_messages = [{"role": "system", "content": system_prompt}]
    for msg in messages:
        full_messages.append({"role": "user", "content": msg})
    
    response = llm.create_chat_completion(
        messages=full_messages,
        temperature=0.7,
        max_tokens=1024
    )
    return response["choices"][0]["message"]["content"]

iface = Interface(
    fn=generate,
    inputs=["text", "text"],
    outputs="text",
    title="Hyperpixel AI Model"
)

iface.launch()