Hugging Face's logo

Spaces:
Rozinamax
/
Text_to_Speech
Sleeping

App Files Community
Text_to_Speech
File size: 2,483 Bytes 
97e4faf
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
deb14ad
e9c4729
 
97e4faf
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
deb14ad
 
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
 
import os
import torch
import gradio as gr
import torchaudio
import time
from datetime import datetime
from tortoise.api import TextToSpeech
from tortoise.utils.text import split_and_recombine_text
from tortoise.utils.audio import load_audio, load_voice, load_voices

VOICE_OPTIONS = [
    "angie",
    "deniro",
    "freeman",
    "random",  # special option for random voice
]

def inference(
    text,
    script,
    voice,
    voice_b,
    seed,
    split_by_newline,
):
    if text is None or text.strip() == "":
        with open(script.name) as f:
            text = f.read()
        if text.strip() == "":
            raise gr.Error("Please provide either text or script file with content.")

    if split_by_newline == "Yes":
        texts = list(filter(lambda x: x.strip() != "", text.split("\n")))
    else:
        texts = split_and_recombine_text(text)

    voices = [voice]
    if voice_b != "disabled":
        voices.append(voice_b)

    if len(voices) == 1:
        voice_samples, conditioning_latents = load_voice(voice)
    else:
        voice_samples, conditioning_latents = load_voices(voices)

    start_time = time.time()

    for j, text in enumerate(texts):
        for audio_frame in tts.tts_with_preset(
            text,
            voice_samples=voice_samples,
            conditioning_latents=conditioning_latents,
            preset="ultra_fast",
            k=1
        ):
            yield (24000, audio_frame.cpu().detach().numpy())

def main():
    title = "Tortoise TTS "
    
    label="Text (Provide either text, or upload a newline separated text file below):",
    )
    script = gr.File(label="Upload a text file")

    voice = gr.Dropdown(
        VOICE_OPTIONS, value="jane_eyre", label="Select voice:", type="value"
    )
    voice_b = gr.Dropdown(
        VOICE_OPTIONS,
        value="disabled",
        label="(Optional) Select second voice:",
        type="value",
    )
    split_by_newline = gr.Radio(
        ["Yes", "No"],
        label="Split by newline (If [No], it will automatically try to find relevant splits):",
        type="value",
        value="No",
    )

    output_audio = gr.Audio(label="streaming audio:", streaming=True, autoplay=True)
    interface = gr.Interface(
        fn=inference,
        inputs=[
            text,
            script,
            voice,
            voice_b,
            split_by_newline,
        ],
        title=title,
        outputs=[output_audio],
    )
    interface.queue().launch()