File size: 3,701 Bytes
e613cea
 
6072bc6
30e6a40
 
 
 
 
 
 
 
 
 
 
dd17486
30e6a40
dd17486
30e6a40
 
 
 
 
 
 
 
 
e227a3a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
30e6a40
9376c53
 
 
365ec8b
9376c53
 
 
365ec8b
9376c53
 
 
 
365ec8b
9376c53
365ec8b
9376c53
 
 
30e6a40
 
 
 
 
 
9376c53
 
 
 
30e6a40
 
 
 
 
 
 
 
 
 
9376c53
 
 
 
 
 
 
 
 
 
 
 
 
 
30e6a40
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
import os
import io
import gradio as gr
import librosa
import numpy as np
import logging
import soundfile
import asyncio
import argparse
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 vc_fn(input_audio, vc_transform, auto_f0):
    if input_audio is None:
        return "You need to upload an audio", None
    sampling_rate, audio = input_audio
    duration = audio.shape[0] / sampling_rate
    if duration > 20 and limitation:
        return "Please upload an audio file that is less than 20 seconds. If you need to generate a longer audio file, please use Colab.", None
    audio = (audio / np.iinfo(audio.dtype).max).astype(np.float32)
    if len(audio.shape) > 1:
        audio = librosa.to_mono(audio.transpose(1, 0))
    if sampling_rate != 16000:
        audio = librosa.resample(audio, orig_sr=sampling_rate, target_sr=16000)
    raw_path = io.BytesIO()
    soundfile.write(raw_path, audio, 16000, format="wav")
    raw_path.seek(0)
    out_audio, out_sr = model.infer(sid, vc_transform, raw_path,
                                   auto_predict_f0=auto_f0,
                                   )
    return "Success", (44100, out_audio.cpu().numpy())

def get_speakers():
  speakers = []

  for _,dirs,_ in os.walk("/models"):
    for folder in dirs:
      cur_speaker = {}
      # Look for G_****.pth
      g = glob.glob(os.path.join("/models",folder,'G_*.pth'))
      if not len(g):
        continue
      cur_speaker["model_path"] = g[0]
      cur_speaker["model_folder"] = folder
      cur_speaker["name"] = folder

      speakers.append(copy.copy(cur_speaker))

  return sorted(speakers, key=lambda x:x["name"].lower())

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")
    args = parser.parse_args()

    speakers = get_speakers()
    speaker_list = [x["name"] for x in speakers]

    models = []
    voices = []

    # !svc infer {NAME}.wav -c config.json -m G_riri_220.pth
    #  display(Audio(f"{NAME}.out.wav", autoplay=True))
    with gr.Blocks() as app:
        gr.Markdown(
            "# <center> Sovits Chapay\n"
            "## <center> The input audio should be clean and pure voice without background music.\n"
        )

        with gr.Row():
            with gr.Column():
                vc_input = gr.Audio(label="Input audio"+' (less than 20 seconds)' if limitation else '')

                vc_transform = gr.Number(label="vc_transform", value=0)

                voice = gr.Dropdown(choices=speaker_list, visible=True)

                vc_submit = gr.Button("Generate", variant="primary")
            with gr.Column():
                vc_output1 = gr.Textbox(label="Output Message")
                vc_output2 = gr.Audio(label="Output Audio")
            vc_submit.click(vc_fn, [vc_input, vc_transform, auto_f0], [vc_output1, vc_output2])
        app.queue(concurrency_count=1, api_open=args.api).launch(share=args.share)