app.py

import gradio as gr
import spaces
from transformers import pipeline

# Create the text generation pipeline.
# If you're running on GPU, you can specify device=0 (or use device_map="auto" if supported).
pipe = pipeline("text-generation", model="TheBloke/Chronoboros-33B-GPTQ", device=0)

@spaces.GPU
def respond(message, history: list[tuple[str, str]], system_message, max_tokens, temperature, top_p):
    # Build the prompt from system message and conversation history.
    prompt = f"{system_message}\n"
    for user_text, assistant_text in history:
        if user_text:
            prompt += f"User: {user_text}\n"
        if assistant_text:
            prompt += f"Assistant: {assistant_text}\n"
    prompt += f"User: {message}\nAssistant: "
    
    # Generate a response using the pipeline.
    # The pipeline returns a list of dictionaries; we take the generated text from the first output.
    output = pipe(prompt, max_new_tokens=max_tokens, temperature=temperature, top_p=top_p)
    full_text = output[0]["generated_text"]
    
    # Remove the prompt from the generated text to isolate the response.
    response_text = full_text[len(prompt):]
    
    # Simulate streaming output in chunks (e.g., 5 characters at a time).
    chunk_size = 5
    for i in range(0, len(response_text), chunk_size):
        yield response_text[: i + chunk_size]

# Configure the ChatInterface with additional inputs.
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
    ],
)

if __name__ == "__main__":
    demo.launch()