import os import gradio as gr import edge_tts from pathlib import Path import inference.infer_tool as infer_tool import utils from inference.infer_tool import Svc import logging import webbrowser import argparse import asyncio import librosa import soundfile import gradio.processing_utils as gr_processing_utils logging.getLogger('numba').setLevel(logging.WARNING) logging.getLogger('markdown_it').setLevel(logging.WARNING) logging.getLogger('urllib3').setLevel(logging.WARNING) logging.getLogger('matplotlib').setLevel(logging.WARNING) limitation = os.getenv("SYSTEM") == "spaces" # limit audio length in huggingface spaces audio_postprocess_ori = gr.Audio.postprocess def audio_postprocess(self, y): data = audio_postprocess_ori(self, y) if data is None: return None return gr_processing_utils.encode_url_or_file_to_base64(data["name"]) gr.Audio.postprocess = audio_postprocess def create_vc_fn(model, sid): def vc_fn(input_audio, vc_transform, auto_f0, slice_db, noise_scale, pad_seconds, tts_text, tts_voice, tts_mode): if tts_mode: if len(tts_text) > 100 and limitation: return "Text is too long", None if tts_text is None or tts_voice is None: return "You need to enter text and select a voice", None asyncio.run(edge_tts.Communicate(tts_text, "-".join(tts_voice.split('-')[:-1])).save("tts.mp3")) audio, sr = librosa.load("tts.mp3") soundfile.write("tts.wav", audio, 24000, format="wav") wav_path = "tts.wav" else: if input_audio is None: return "You need to select an audio", None raw_audio_path = f"raw/{input_audio}" if "." not in raw_audio_path: raw_audio_path += ".wav" infer_tool.format_wav(raw_audio_path) wav_path = Path(raw_audio_path).with_suffix('.wav') _audio = model.slice_inference( wav_path, sid, vc_transform, slice_db, cluster_infer_ratio=0, auto_predict_f0=auto_f0, noice_scale=noise_scale, pad_seconds=pad_seconds) model.clear_empty() return "Success", (44100, _audio) return vc_fn def refresh_raw_wav(): return gr.Dropdown.update(choices=os.listdir("raw")) def change_to_tts_mode(tts_mode): if tts_mode: return gr.Audio.update(visible=False), gr.Button.update(visible=False), gr.Textbox.update(visible=True), gr.Dropdown.update(visible=True) else: return gr.Audio.update(visible=True), gr.Button.update(visible=True), gr.Textbox.update(visible=False), gr.Dropdown.update(visible=False) if __name__ == '__main__': parser = argparse.ArgumentParser() parser.add_argument('--device', type=str, default='cpu') parser.add_argument('--api', action="store_true", default=False) parser.add_argument("--share", action="store_true", default=False, help="share gradio app") parser.add_argument("--colab", action="store_true", default=False, help="share gradio app") args = parser.parse_args() hubert_model = utils.get_hubert_model().to(args.device) models = [] voices = [] tts_voice_list = asyncio.get_event_loop().run_until_complete(edge_tts.list_voices()) for r in tts_voice_list: voices.append(f"{r['ShortName']}-{r['Gender']}") raw = os.listdir("raw") for f in os.listdir("models"): name = f model = Svc(fr"models/{f}/{f}.pth", f"models/{f}/config.json", device=args.device, hubert_model=hubert_model) cover = f"models/{f}/cover.png" if os.path.exists(f"models/{f}/cover.png") else None models.append((name, cover, create_vc_fn(model, name))) with gr.Blocks() as app: gr.Markdown( "#