File size: 3,626 Bytes
2337a8e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
from typing import Dict, Optional, Union

import numpy as np

from .generation import codec_decode, generate_coarse, generate_fine, generate_text_semantic


def text_to_semantic(
    text: str,
    history_prompt: Optional[Union[Dict, str]] = None,
    temp: float = 0.7,
    silent: bool = False,
):
    """Generate semantic array from text.

    Args:
        text: text to be turned into audio
        history_prompt: history choice for audio cloning
        temp: generation temperature (1.0 more diverse, 0.0 more conservative)
        silent: disable progress bar

    Returns:
        numpy semantic array to be fed into `semantic_to_waveform`
    """
    x_semantic = generate_text_semantic(
        text,
        history_prompt=history_prompt,
        temp=temp,
        silent=silent,
        use_kv_caching=True
    )
    return x_semantic


def semantic_to_waveform(
    semantic_tokens: np.ndarray,
    history_prompt: Optional[Union[Dict, str]] = None,
    temp: float = 0.7,
    silent: bool = False,
    output_full: bool = False,
):
    """Generate audio array from semantic input.

    Args:
        semantic_tokens: semantic token output from `text_to_semantic`
        history_prompt: history choice for audio cloning
        temp: generation temperature (1.0 more diverse, 0.0 more conservative)
        silent: disable progress bar
        output_full: return full generation to be used as a history prompt

    Returns:
        numpy audio array at sample frequency 24khz
    """
    coarse_tokens = generate_coarse(
        semantic_tokens,
        history_prompt=history_prompt,
        temp=temp,
        silent=silent,
        use_kv_caching=True
    )
    fine_tokens = generate_fine(
        coarse_tokens,
        history_prompt=history_prompt,
        temp=0.5,
    )
    audio_arr = codec_decode(fine_tokens)
    if output_full:
        full_generation = {
            "semantic_prompt": semantic_tokens,
            "coarse_prompt": coarse_tokens,
            "fine_prompt": fine_tokens,
        }
        return full_generation, audio_arr
    return audio_arr


def save_as_prompt(filepath, full_generation):
    assert(filepath.endswith(".npz"))
    assert(isinstance(full_generation, dict))
    assert("semantic_prompt" in full_generation)
    assert("coarse_prompt" in full_generation)
    assert("fine_prompt" in full_generation)
    np.savez(filepath, **full_generation)


def generate_audio(
    text: str,
    history_prompt: Optional[Union[Dict, str]] = None,
    text_temp: float = 0.7,
    waveform_temp: float = 0.7,
    silent: bool = False,
    output_full: bool = False,
):
    """Generate audio array from input text.

    Args:
        text: text to be turned into audio
        history_prompt: history choice for audio cloning
        text_temp: generation temperature (1.0 more diverse, 0.0 more conservative)
        waveform_temp: generation temperature (1.0 more diverse, 0.0 more conservative)
        silent: disable progress bar
        output_full: return full generation to be used as a history prompt

    Returns:
        numpy audio array at sample frequency 24khz
    """
    semantic_tokens = text_to_semantic(
        text,
        history_prompt=history_prompt,
        temp=text_temp,
        silent=silent,
    )
    out = semantic_to_waveform(
        semantic_tokens,
        history_prompt=history_prompt,
        temp=waveform_temp,
        silent=silent,
        output_full=output_full,
    )
    if output_full:
        full_generation, audio_arr = out
        return full_generation, audio_arr
    else:
        audio_arr = out
    return audio_arr