Hugging Face's logo

Spaces:
AkitoP
/
GPT-SoVITS-V2-misc_models
Running on Zero

App Files Community
1
GPT-SoVITS-V2-misc_models
File size: 2,976 Bytes 
4536f60
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
import time

import librosa
import torch
import torch.nn.functional as F
import soundfile as sf
import logging

logging.getLogger("numba").setLevel(logging.WARNING)

from transformers import (
    Wav2Vec2FeatureExtractor,
    HubertModel,
)

import torch.nn as nn

cnhubert_base_path = None


class CNHubert(nn.Module):
    def __init__(self):
        super().__init__()
        self.model = HubertModel.from_pretrained(cnhubert_base_path)
        self.feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained(
            cnhubert_base_path
        )

    def forward(self, x):
        input_values = self.feature_extractor(
            x, return_tensors="pt", sampling_rate=16000
        ).input_values.to(x.device)
        feats = self.model(input_values)["last_hidden_state"]
        return feats


# class CNHubertLarge(nn.Module):
#     def __init__(self):
#         super().__init__()
#         self.model = HubertModel.from_pretrained("/data/docker/liujing04/gpt-vits/chinese-hubert-large")
#         self.feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained("/data/docker/liujing04/gpt-vits/chinese-hubert-large")
#     def forward(self, x):
#         input_values = self.feature_extractor(x, return_tensors="pt", sampling_rate=16000).input_values.to(x.device)
#         feats = self.model(input_values)["last_hidden_state"]
#         return feats
#
# class CVec(nn.Module):
#     def __init__(self):
#         super().__init__()
#         self.model = HubertModel.from_pretrained("/data/docker/liujing04/vc-webui-big/hubert_base")
#         self.feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained("/data/docker/liujing04/vc-webui-big/hubert_base")
#     def forward(self, x):
#         input_values = self.feature_extractor(x, return_tensors="pt", sampling_rate=16000).input_values.to(x.device)
#         feats = self.model(input_values)["last_hidden_state"]
#         return feats
#
# class cnw2v2base(nn.Module):
#     def __init__(self):
#         super().__init__()
#         self.model = Wav2Vec2Model.from_pretrained("/data/docker/liujing04/gpt-vits/chinese-wav2vec2-base")
#         self.feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained("/data/docker/liujing04/gpt-vits/chinese-wav2vec2-base")
#     def forward(self, x):
#         input_values = self.feature_extractor(x, return_tensors="pt", sampling_rate=16000).input_values.to(x.device)
#         feats = self.model(input_values)["last_hidden_state"]
#         return feats


def get_model():
    model = CNHubert()
    model.eval()
    return model


# def get_large_model():
#     model = CNHubertLarge()
#     model.eval()
#     return model
#
# def get_model_cvec():
#     model = CVec()
#     model.eval()
#     return model
#
# def get_model_cnw2v2base():
#     model = cnw2v2base()
#     model.eval()
#     return model


def get_content(hmodel, wav_16k_tensor):
    with torch.no_grad():
        feats = hmodel(wav_16k_tensor)
    return feats.transpose(1, 2)