Hugging Face's logo

rippertnt
/
on-vits2-multi-tts-v1

ONNX
Model card Files Community
on-vits2-multi-tts-v1
File size: 2,696 Bytes 
8d763f2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
## VCTK
import torch

import commons
import utils
from models import SynthesizerTrn
from text.symbols import symbols
from text import text_to_sequence
import time 

from scipy.io.wavfile import write


def get_text(text, hps):
    text_norm = text_to_sequence(text, hps.data.text_cleaners)
    if hps.data.add_blank:
        text_norm = commons.intersperse(text_norm, 0)
    print(text, text_norm)
    text_norm = torch.LongTensor(text_norm)
    return text_norm

LANG = 'ru'
CONFIG_PATH = f"./configs/{LANG}_base.json"
MODEL_PATH = f"./logs/{LANG}_base/G_40000.pth"
#TEXT = "I am artificial intelligent voice made by circulus."
#TEXT = "저는 서큘러스의 AI Voice 모델입니다. 오늘도 즐거운하루 보내세요."
TEXT = "привет. Я президент Путин, и мне нравятся советские лидеры Сталин и Ленин."
#TEXT = "Xin chào. Tôi là Tổng thống Putin và tôi thích các nhà lãnh đạo Liên Xô Stalin và Lenin."
#TEXT = "สวัสดี. ผมเป็นประธานาธิบดีปูติน และผมชอบผู้นำโซเวียตอย่างสตาลินและเลนิน"
#TEXT = "Halo. Saya Presiden Putin, dan saya menyukai pemimpin Soviet Stalin dan Lenin."

hps = utils.get_hparams_from_file(CONFIG_PATH)

if (
    "use_mel_posterior_encoder" in hps.model.keys()
    and hps.model.use_mel_posterior_encoder == True
):
    print("Using mel posterior encoder for VITS2")
    posterior_channels = 80  # vits2
    hps.data.use_mel_posterior_encoder = True
else:
    print("Using lin posterior encoder for VITS1")
    posterior_channels = hps.data.filter_length // 2 + 1
    hps.data.use_mel_posterior_encoder = False

net_g = SynthesizerTrn(
    len(symbols),
    posterior_channels,
    hps.train.segment_size // hps.data.hop_length,
    n_speakers=hps.data.n_speakers,
    **hps.model
)
_ = net_g.eval()

_ = utils.load_checkpoint(MODEL_PATH, net_g, None)

stn_tst = get_text(TEXT, hps)

with torch.no_grad():

    for i in range(0,hps.data.n_speakers):
        start = time.time()
        x_tst = stn_tst.unsqueeze(0)
        x_tst_lengths = torch.LongTensor([stn_tst.size(0)])
        sid = torch.LongTensor([i])
        audio = (
            net_g.infer(
                x_tst,
                x_tst_lengths,
                sid=sid,
                noise_scale=0.667,
                noise_scale_w=0.8,
                length_scale=1,
            )[0][0, 0]
            .data
            .float()
            .numpy()
        )
        print(i, time.time() - start)
        write(data=audio, rate=hps.data.sampling_rate, filename=f"test_{LANG}_{i}.wav")