Spaces:
Running
Running
| import gradio as gr | |
| import os | |
| import requests | |
| import torch | |
| import zipfile | |
| from TTS.api import TTS | |
| from pydub import AudioSegment | |
| #download for mecab | |
| os.system('python -m unidic download') | |
| os.environ["COQUI_TOS_AGREED"] = "1" | |
| MODEL_PATH = "tts_models/multilingual/multi-dataset/xtts_v2" | |
| LANGUAGES = ["en", "es", "fr", "de", "it", "pt", "pl", "tr", "ru", "nl", "cs", "ar", "zh-cn", "ja", "hu", "ko", "hi"] | |
| AUDIO_FORMATS = [".wav", ".mp3", ".flac", ".mp4"] | |
| device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu") | |
| print(f"Using device: {device}") | |
| tts = TTS(MODEL_PATH).to(device) | |
| def download_audio_file(url): | |
| try: | |
| response = requests.get(url, stream=True) | |
| file_name = url.split("/")[-1] | |
| file_extension = os.path.splitext(file_name)[-1].lower() | |
| if file_extension not in AUDIO_FORMATS: | |
| raise ValueError(f"Invalid audio file format: {file_extension}") | |
| with open(file_name, "wb") as f: | |
| f.write(response.content) # Write the entire response content at once | |
| return file_name | |
| except requests.exceptions.RequestException as e: | |
| print(f"Error downloading audio file: {e}") | |
| return None | |
| def convert_to_wav(input_audio_file): | |
| file_extension = os.path.splitext(input_audio_file)[-1].lower() | |
| if file_extension!= ".wav": | |
| audio = AudioSegment.from_file(input_audio_file) | |
| audio.export("temp.wav", format="wav") | |
| os.remove(input_audio_file) | |
| return "temp.wav" | |
| return input_audio_file | |
| def synthesize_text(text, input_audio_file, language): | |
| input_audio_file = convert_to_wav(input_audio_file) | |
| tts.tts_to_file(text=text, speaker_wav=input_audio_file, language=language, file_path="./output.wav") | |
| return "./output.wav" | |
| def clone(text, input_file, language, url=None, use_url=False): | |
| if use_url: | |
| if url is None: | |
| return None | |
| input_audio_file = download_audio_file(url) | |
| if input_audio_file is None: | |
| return None | |
| else: | |
| if input_file is None: | |
| return None | |
| input_audio_file = input_file.name | |
| output_file_path = synthesize_text(text, input_audio_file, language) | |
| return output_file_path | |
| iface = gr.Interface( | |
| fn=clone, | |
| inputs=["text", gr.File(label="Input File", file_types=AUDIO_FORMATS), gr.Dropdown(choices=LANGUAGES, label="Language"), gr.Text(label="URL"), gr.Checkbox(label="Use URL", value=False)], | |
| outputs=gr.Audio(type='filepath'), | |
| title='Voice Clone', | |
| description=""" by [Angetyde](https://youtube.com/@Angetyde?si=7nusP31nTumIkPTF) and [Tony Assi](https://www.tonyassi.com/ ) use this colab with caution <3. """, | |
| theme=gr.themes.Base(primary_hue="teal", secondary_hue="teal", neutral_hue="slate") | |
| ) | |
| iface.launch(share=True) |