app.py

import gradio as gr
import pyttsx3
import torch
from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer
import whisper

# Load text-generation model
model_name = "microsoft/DialoGPT-medium"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name)

# Initialize Text-to-Speech
tts = pyttsx3.init()
tts.setProperty("rate", 150)  # Adjust speech speed

# Load Whisper for speech recognition
whisper_model = whisper.load_model("small")

def live_chat(audio):
    # Convert speech to text
    result = whisper_model.transcribe(audio)
    user_input = result["text"]

    # Generate chatbot response
    input_ids = tokenizer.encode(user_input + tokenizer.eos_token, return_tensors="pt")
    response_ids = model.generate(input_ids, max_length=100, pad_token_id=tokenizer.eos_token_id)
    bot_response = tokenizer.decode(response_ids[:, input_ids.shape[-1]:][0], skip_special_tokens=True)

    # Speak response
    tts.say(bot_response)
    tts.runAndWait()

    return bot_response

# Gradio Interface
app = gr.Interface(
    fn=live_chat,
    inputs=gr.Audio(type="filepath"),  # Microphone input
    outputs="text",
    live=True,
)

# Launch the app
app.launch()