File size: 4,132 Bytes
0019e18
 
1e50de0
0019e18
 
84e9bbc
2873a94
 
0019e18
 
a7e530a
 
2873a94
0019e18
 
 
 
 
 
2873a94
0019e18
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
84e9bbc
0019e18
 
 
 
 
 
 
 
 
 
 
 
b31036b
 
0019e18
 
 
84e9bbc
0019e18
 
 
 
 
 
 
 
 
 
 
 
 
 
a7e530a
0019e18
 
 
 
 
 
 
 
 
 
 
 
1e50de0
2873a94
1e50de0
0019e18
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2873a94
0019e18
 
 
 
 
 
 
 
 
 
2238e1d
0019e18
 
 
 
 
f0aa454
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
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
import gradio as gr
from infer_rvc_python import BaseLoader
import soundfile as sf
import random
from urllib.request import urlretrieve
import os
import zipfile

files_to_retrieve = [
    "https://replicate.delivery/pbxt/N97QM3XNFrooJhV6Fb0meBff0aAG1rEDfvuxcdLS6fTx1vmWC/test.zip",
    # "https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/hubert_base.pt?download=true",
    # "https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/rmvpe.pt?download=true"
]

for file in files_to_retrieve:
    print(f"Downloading {file}")
    urlretrieve(file, file.split("/")[-1])

# unzip test.zip
with zipfile.ZipFile("test.zip", "r") as zip_ref:
    zip_ref.extractall(".")


converter = BaseLoader(
    only_cpu=True, hubert_path="./hubert_base.pt", rmvpe_path="./rmvpe.pt"
)

model = "test.pth"
index = "added_IVF839_Flat_nprobe_1_test_v2.index"


def voice_conversion(
    audio,
    pitch_change,
    filter_radius,
    envelope_ratio,
    index_influence,
    consonant_breath_protection,
):
    global output_file
    audio_out = run(
        [str(audio)],
        model,
        "rmvpe+",
        pitch_change,
        index,
        index_influence,
        filter_radius,
        envelope_ratio,
        consonant_breath_protection,
    )
    print(audio_out)
    # output_audio, sr = sf.read(output_file, dtype="int32")
    return audio_out


def convert_now(audio_files, random_tag):
    return converter(audio_files, random_tag, overwrite=False, parallel_workers=8)


def run(
    audio_files,
    file_m,
    pitch_alg,
    pitch_lvl,
    file_index,
    index_inf,
    r_m_f,
    e_r,
    c_b_p,
):
    random_tag = "USER_" + str(random.randint(10000000, 99999999))
    print("PITCH LVL: ", pitch_lvl)
    converter.apply_conf(
        tag=random_tag,
        file_model=file_m,
        pitch_algo=pitch_alg,
        pitch_lvl=pitch_lvl,
        file_index=file_index,
        index_influence=index_inf,
        respiration_median_filtering=r_m_f,
        envelope_ratio=e_r,
        consonant_breath_protection=c_b_p,
        resample_sr=44100 if audio_files[0].endswith(".mp3") else 0,
    )
    output = convert_now(audio_files, random_tag)
    audio, sr = sf.read(output[0], dtype="int32")
    return (sr, audio)


def ui():
    with gr.Blocks() as demo:
        audio_input = gr.Audio(sources=["microphone", "upload"], type="filepath")
        with gr.Row():
            pitch_slider = gr.Slider(
                minimum=-24,
                maximum=24,
                value=0,
                step=1,
                label="Pitch",
                interactive=True,
            )
            index_influence_slider = gr.Slider(
                minimum=0,
                maximum=1,
                value=0.75,
                step=0.01,
                label="Index Influence",
                interactive=True,
            )
            respiration_median_filtering = gr.Slider(
                minimum=0,
                maximum=10,
                value=3,
                step=1,
                label="Resp. Median Filtering",
                interactive=True,
            )
            envelope_ratio = gr.Slider(
                minimum=0,
                maximum=1,
                value=0.25,
                step=0.01,
                label="Envelope Ratio",
                interactive=True,
            )
            consonant_breath_protection = gr.Slider(
                minimum=0,
                maximum=1,
                value=0.5,
                step=0.01,
                label="Consonant Breath Protection",
                interactive=True,
            )
        button = gr.Button("Convert")
        audio_output = gr.Audio(interactive=False, type="numpy")
        button.click(
            voice_conversion,
            inputs=[
                audio_input,
                pitch_slider,
                respiration_median_filtering,
                envelope_ratio,
                index_influence_slider,
                consonant_breath_protection,
            ],
            outputs=[audio_output],
        )

    return demo


ui().launch(auth=("output", "becreative"))