File size: 3,498 Bytes
9659078
5f924a4
 
9659078
adf3e3f
f523090
 
adf3e3f
5a84593
 
 
bd64281
5a84593
 
 
adf3e3f
65814fd
 
9659078
589b334
65814fd
 
 
9659078
5a82610
 
 
5f924a4
 
8302c0f
 
5f924a4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5a82610
e236784
9659078
8302c0f
9659078
a4b64d4
 
 
 
0435c60
 
 
 
ed4738c
 
0435c60
 
8e00ffa
9659078
5a82610
 
3c31edb
8302c0f
0435c60
 
 
 
 
5a82610
 
 
 
 
 
f523090
0c32eee
 
 
 
 
 
 
adf3e3f
8302c0f
0cdadc9
0bd14d2
 
 
 
 
 
 
 
 
 
 
 
0435c60
0bd14d2
 
 
0435c60
 
 
 
0bd14d2
 
0cdadc9
0bd14d2
 
0435c60
0bd14d2
 
 
0435c60
 
0bd14d2
 
 
 
 
 
 
 
 
adf3e3f
0bd14d2
 
 
 
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
import gradio as gr
import os 
import shutil

#from huggingface_hub import snapshot_download
import numpy as np
from scipy.io import wavfile
"""
model_ids = [
    'suno/bark',
]

for model_id in model_ids:
    model_name = model_id.split('/')[-1]
    snapshot_download(model_id, local_dir=f'checkpoints/{model_name}')
"""
#from TTS.tts.configs.bark_config import BarkConfig
#from TTS.tts.models.bark import Bark

#os.environ['CUDA_VISIBLE_DEVICES'] = '1'
#config = BarkConfig()
#model = Bark.init_from_config(config)
#model.load_checkpoint(config, checkpoint_dir="checkpoints/bark", eval=True)

from TTS.api import TTS
tts = TTS("tts_models/multilingual/multi-dataset/bark", gpu=True)

def infer(prompt, input_wav_file):

    print("SAVING THE AUDIO FILE TO WHERE IT BELONGS")

    # Path to your WAV file
    source_path = input_wav_file

    # Destination directory
    destination_directory = "bark_voices"

    # Extract the file name without the extension
    file_name = os.path.splitext(os.path.basename(source_path))[0]

    # Construct the full destination directory path
    destination_path = os.path.join(destination_directory, file_name)

    # Create the new directory
    os.makedirs(destination_path, exist_ok=True)

    # Move the WAV file to the new directory
    shutil.move(source_path, os.path.join(destination_path, f"{file_name}.wav"))

    """
    text = prompt

    print("SYNTHETIZING...")
    # with random speaker
    #output_dict = model.synthesize(text, config, speaker_id="random", voice_dirs=None)

    # cloning a speaker.
    # It assumes that you have a speaker file in `bark_voices/speaker_n/speaker.wav` or `bark_voices/speaker_n/speaker.npz`
    output_dict = model.synthesize(
        text, 
        config, 
        speaker_id=f"{file_name}", 
        voice_dirs="bark_voices/",
        gpu=True
    )
    
    print(output_dict)

    

    sample_rate = 24000  # Replace with the actual sample rate
    print("WRITING WAVE FILE")
    wavfile.write(
        'output.wav', 
        sample_rate, 
        output_dict['wav']
    )
    """
    
    tts.tts_to_file(text=prompt,
                file_path="output.wav",
                voice_dir="bark_voices/",
                speaker=f"{file_name}")

    # List all the files and subdirectories in the given directory
    contents = os.listdir(f"bark_voices/{file_name}")

    # Print the contents
    for item in contents:
        print(item)   
    
    return "output.wav", f"bark_voices/{file_name}/{contents[1]}"


css = """
#col-container {max-width: 780px; margin-left: auto; margin-right: auto;}
"""

with gr.Blocks(css=css) as demo:
    with gr.Column(elem_id="col-container"):
        
        gr.HTML("""
        <h1>Instant Voice Cloning</h1>
        """)
        
        prompt = gr.Textbox(
            label="Text to speech prompt"
        )
        
        audio_in = gr.Audio(
            label="WAV voice to clone", 
            type="filepath",
            source="upload"
        )
        
        submit_btn = gr.Button("Submit")
        submit_with_npz_btn = gr.Button("Submit 2", visible=False)
        
        cloned_out = gr.Audio(
            label="Text to speech output"
        )
        
        npz_file = gr.File(
            label=".npz file"
        )
    
    submit_btn.click(
        fn = infer,
        inputs = [
            prompt,
            audio_in
        ],
        outputs = [
            cloned_out, 
            npz_file
        ]
    )

demo.queue().launch()