app.py

import gradio as gr
import os
from transformers import pipeline, AutoTokenizer

# Load the tokenizer and model using the pipeline
pipe = pipeline("text-generation", model="explorewithai/Loxa-4B", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("explorewithai/Loxa-4B")

# Get the system prompt from environment variables
meo_system = os.environ.get("MEO")

def respond(
    message,
    history,
    max_tokens,
    temperature,
    top_p,
):
    # Format the messages for the pipeline
    messages = [{"role": "system", "content": meo_system}]
    for user_msg, bot_msg in history:
        messages.append({"role": "user", "content": user_msg})
        messages.append({"role": "assistant", "content": bot_msg})
    messages.append({"role": "user", "content": message})

    # Generate the prompt using the tokenizer's chat template
    prompt = tokenizer.apply_chat_template(messages, tokenize=False)

    # Generate the response using the pipeline
    outputs = pipe(
        prompt,
        max_new_tokens=max_tokens,
        do_sample=True,
        temperature=temperature,
        top_p=top_p,
        return_full_text=False  # We only want the generated part
    )

    # Extract the generated text
    response = outputs[0]['generated_text']

    return response

# Create the Gradio interface
demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=1.0, 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()