app.py

from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
from transformers import MarianMTModel, MarianTokenizer
from elevenlabs import set_api_key
from elevenlabs import clone, generate, play
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
from pydub import AudioSegment
from huggingface_hub.hf_api import HfFolder
HfFolder.save_token('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')
access_token = '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'

import streamlit as st


#language translation function 
def translate_english_to_hindi(input_text):
    # Load the pre-trained English to Hindi translation model and tokenizer
    model_name = "Helsinki-NLP/opus-mt-en-hi"
    tokenizer = MarianTokenizer.from_pretrained(model_name)
    model = MarianMTModel.from_pretrained(model_name)

    # Tokenize the input text and generate translation
    inputs = tokenizer(input_text, return_tensors="pt", padding=True)
    translated_ids = model.generate(inputs.input_ids)

    # Decode the translated output
    translated_text = tokenizer.decode(translated_ids[0], skip_special_tokens=True)

    return translated_text


def save_uploaded_file_as_mp3(uploaded_file, output_file_path):
    audio = AudioSegment.from_file(uploaded_file)
    audio.export(output_file_path, format="mp3")

def ui(): 
    st.title('Multi Lingual Voice Cloner')
    st.markdown("Made by vageesh")

    #audio input box 
    audio_file = st.file_uploader("Upload an audio file that needs to be cloned", type=[ "wav,Mp4","Mp3"])
    if audio_file is not None:
        output_file_path = "./output_audio.mp3"
        save_uploaded_file_as_mp3(audio_file, output_file_path)
        st.success(f"Audio file saved as {output_file_path}")
        user_api_key = st.sidebar.text_input(
            label="#### Your Eleven Labs API key here 👇",
            placeholder="Paste your Eleven Labs API key API key, sk-",
            type="password")
    
        if user_api_key is not None and user_api_key.strip() != "":
            set_api_key(user_api_key)
            #making an voice 
            voice = clone(
                name="Alex",
                description="An middle aged American male voice with a slight hoarseness in his throat. Perfect for npodcast", # Optional
                files=["./output_audio.mp3"],
                model='eleven_multilingual_v1',
                language='Hindi'
            )

            in_text=st.text_input("Paste the text you want to hear from english to hindi")
            if in_text is not None and in_text.strip() != "":
                hin_text=translate_english_to_hindi(in_text)
                audio = generate(text=hin_text, voice=voice)

                st.audio(audio)

if __name__=="__main__": 
    ui()