Spaces:
Running
Running
| import subprocess | |
| import os | |
| import torch | |
| import urllib.request | |
| import librosa | |
| from moviepy.editor import VideoFileClip | |
| from TTS.api import TTS | |
| # Check if PyTorch is using the GPU for computations | |
| if torch.cuda.is_available(): | |
| device = torch.device("cuda") | |
| print("Using the GPU for computations") | |
| else: | |
| device = torch.device("cpu") | |
| print("Using the CPU for computations") | |
| os.environ["COQUI_TOS_AGREED"] = "1" | |
| tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2").to(device) | |
| def convert_audio_to_wav(file_path): | |
| """Convert the given audio file to WAV format.""" | |
| file_name = os.path.basename(file_path) | |
| file_ext = os.path.splitext(file_name)[1].lower() | |
| if file_ext == ".mp3": | |
| audio, sr = librosa.load(file_path) | |
| librosa.output.write_wav(f"temp_{file_name}", audio, sr) | |
| file_path = f"temp_{file_name}" | |
| elif file_ext == ".flac": | |
| os.system(f"ffmpeg -i {file_path} -acodec pcm_s16le -ar 16000 temp_{file_name}") | |
| file_path = f"temp_{file_name}" | |
| elif file_ext == ".mp4": | |
| clip = VideoFileClip(file_path, audio_codec="aac") | |
| audio = clip.audio | |
| audio.write_audiofile(f"temp_{file_name}") | |
| file_path = f"temp_{file_name}" | |
| return file_path | |
| def clone(text, url, language): | |
| """Generate a voice clone using the given parameters.""" | |
| response = requests.get(url) | |
| with open("temp.zip", "wb") as f: | |
| f.write(response.content) | |
| with zipfile.ZipFile("temp.zip", "r") as zip_ref: | |
| zip_ref.extractall() | |
| audio_file = [f for f in os.listdir(".") if f.endswith(".wav")][0] | |
| # Convert the audio file to WAV format | |
| if os.path.splitext(audio_file)[1].lower() not in [".wav", ".flac"]: | |
| audio_file = convert_audio_to_wav(audio_file) | |
| # Check if a GPU is available | |
| if torch.cuda.is_available(): | |
| # Set the device to the GPU | |
| device = torch.device("cuda") | |
| print("Using the GPU for computations") | |
| else: | |
| # Set the device to the CPU | |
| device = torch.device("cpu") | |
| print("Using the CPU for computations") | |
| # Load the TTS model and move it to the device | |
| tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2").to(device) | |
| tts.tts_to_file(text=text, speaker_wav=audio_file, language=language, file_path="./output.wav") | |
| os.remove(audio_file) | |
| os.remove("temp.zip") | |
| return "./output.wav" | |
| iface = gr.Interface(fn=clone, | |
| inputs=["text", gr.components.Text(label="URL"), gr.Dropdown(choices=["en", "es", "fr", "de", "it", "pt", "pl", "tr", "ru", "nl", "cs", "ar", "zh-cn", "ja", "hu", "ko", "hi"], label="Language")], | |
| outputs=gr.Audio(type='filepath'), | |
| title='Voice Clone', | |
| description=""" | |
| by [Angetyde](https://youtube.com/@Angetyde?si=7nusP31nTumIkPTF) and$@$v=v1.16$@$[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) |