app.py

import gradio as gr
from huggingface_hub import InferenceClient
import os
from gtts import gTTS
import whisper
import io
from tempfile import NamedTemporaryFile

api = os.getenv("HF_API_TOKEN")
client = InferenceClient("meta-llama/Meta-Llama-3.1-70B-Instruct", token=f"{api}")

# Load Whisper model
model = whisper.load_model("base")  # or use 'small', 'medium', 'large', depending on your needs

def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    messages = [{"role": "system", "content": system_message}]

    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})

    messages.append({"role": "user", "content": message})

    response = ""

    for message in client.chat_completion(
        messages,
        max_tokens=max_tokens,
        stream=True,
        temperature=temperature,
        top_p=top_p,
    ):
        token = message.choices[0].delta.content

        response += token
        yield response

def text_to_speech(text):
    tts = gTTS(text=text, lang='en')
    with NamedTemporaryFile(delete=True) as tmpfile:
        tts.save(tmpfile.name)
        with open(tmpfile.name, "rb") as f:
            return f.read()

def speech_to_text(audio):
    # Load audio data into a temporary file
    with NamedTemporaryFile(delete=True, suffix=".wav") as tmpfile:
        tmpfile.write(audio)
        tmpfile.flush()
        
        # Transcribe audio with Whisper
        result = model.transcribe(tmpfile.name)
        return result['text']

def process_audio(audio, system_message, max_tokens, temperature, top_p):
    text = speech_to_text(audio)
    response_gen = respond(
        message=text,
        history=[],
        system_message=system_message,
        max_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
    )
    response_text = next(response_gen)
    audio_response = text_to_speech(response_text)
    return audio_response

demo = gr.Interface(
    fn=process_audio,
    inputs=[
        gr.Audio(source="microphone", type="bytes"),
        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)",
        ),
    ],
    outputs=gr.Audio(type="bytes"),
)

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