File size: 4,436 Bytes
afd85cb
 
 
 
 
4f17706
afd85cb
 
 
 
2ea2c52
afd85cb
 
 
 
33de63c
 
 
 
 
 
 
 
 
 
 
 
f6ffbcf
afd85cb
 
33de63c
afd85cb
33de63c
afd85cb
 
976a73f
33de63c
 
afd85cb
33de63c
afd85cb
 
33de63c
 
c385861
afd85cb
 
 
 
 
 
 
 
 
65b6b11
 
 
 
 
 
 
afd85cb
 
 
2d5ca02
 
 
 
 
 
 
 
 
 
 
 
 
afd85cb
2d5ca02
 
 
afd85cb
2d5ca02
 
afd85cb
2d5ca02
 
 
 
 
 
 
4354e8c
2d5ca02
 
 
33de63c
2d5ca02
 
 
33de63c
2d5ca02
d588216
2d5ca02
 
afd85cb
d8a47b1
afd85cb
d8a47b1
afd85cb
d8a47b1
 
afd85cb
d8a47b1
eb2d1d3
afd85cb
2d5ca02
d8a47b1
afd85cb
 
 
 
65b6b11
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
import streamlit as st
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import json
import os
import tempfile
import shutil
import requests
from pathlib import Path
temp_dir = os.path.expanduser("~/app")
global ckpt_temp_file
global audio_temp_file
global config_temp_file
###################################################
from utils.hparams import hparams
from preprocessing.data_gen_utils import get_pitch_parselmouth,get_pitch_crepe
import numpy as np
import matplotlib.pyplot as plt
import IPython.display as ipd
import utils
import librosa
import torchcrepe
from infer import *
import logging
from infer_tools.infer_tool import *
import io

clip_completed = False
def render_audio(ckpt_temp_file, config_temp_file, audio_temp_file, title):
    logging.getLogger('numba').setLevel(logging.WARNING)
    title = int(title)
    project_name = "Unnamed"
    model_path = ckpt_temp_file
    config_path= config_temp_file
    hubert_gpu=True
    svc_model = Svc(project_name,config_path,hubert_gpu, model_path)
    print('model loaded')
    wav_fn = audio_temp_file
    demoaudio, sr = librosa.load(wav_fn)
    key = title # 音高调整,支持正负(半音)
    # 加速倍数
    pndm_speedup = 20
    wav_gen='queeeeee.wav'#直接改后缀可以保存不同格式音频,如flac可无损压缩
    
    # Show the spinner and run the run_clip function inside the 'with' block
    with st.spinner("Rendering Audio..."):
      f0_tst, f0_pred, audio = run_clip(svc_model,file_path=wav_fn, key=key, acc=pndm_speedup, use_crepe=True, use_pe=True, thre=0.05,
                                        use_gt_mel=False, add_noise_step=500,project_name=project_name,out_path=wav_gen)
    clip_completed = True
    if clip_completed:
    # If the 'run_clip' function has completed, use the st.audio function to show an audio player for the file stored in the 'wav_gen' variable
        st.audio(wav_gen)

#######################################################
st.set_page_config(
    page_title="DiffSVC Render",
    page_icon="🧊",
    initial_sidebar_state="expanded",
)
############
st.title('DIFF-SVC Render')

###CKPT LOADER
with tempfile.TemporaryDirectory(dir=os.path.expanduser("~/app")) as temp_dir:
    ckpt = st.file_uploader("Choose your CKPT", type= 'ckpt')
    # Check if user uploaded a CKPT file
    if ckpt is not None:
      #TEMP FUNCTION
      with tempfile.NamedTemporaryFile(mode="wb", suffix='.ckpt', delete=False, dir=temp_dir) as temp:
        # Get the file contents as bytes
        bytes_data = ckpt.getvalue()
        # Write the bytes to the temporary file
        temp.write(bytes_data)
        ckpt_temp_file = temp.name
        # Print the temporary file name
        print(temp.name)

    # Display the file path
    if "ckpt_temp_file" in locals():
        st.success("File saved to: {}".format(ckpt_temp_file))

    # File uploader
    config = st.file_uploader("Choose your config", type= 'yaml')

    # Check if user uploaded a config file
    if config is not None:
      #TEMP FUNCTION
      with tempfile.NamedTemporaryFile(mode="w", suffix='.yaml', delete=False, dir=temp_dir) as temp:
        # Get the file contents as a string
        str_data = config.read()
        # Write the string to the temporary file
        temp.write(bytes_data.decode())
        config_temp_file = temp.name
        # Print the temporary file name
        print(temp.name)

    # Display the file path
    if "config_temp_file" in locals():
        st.success("File saved to: {}".format(config_temp_file))

    audio = st.file_uploader("Choose your audio", type=["wav", "mp3"])

    # Check if user uploaded an audio file
    if audio is not None:
  #TEMP FUNCTION
      with tempfile.NamedTemporaryFile(mode="wb", suffix='.wav', delete=False, dir=temp_dir) as temp:
    # Get the file contents as bytes
        bytes_data = audio.getvalue()
    # Write the bytes to the temporary file
        temp.write(bytes_data)
        audio_temp_file = temp.name
    # Print the temporary file name
        print(temp.name)

# Display the file path
    if "audio_temp_file" in locals():
        st.success("File saved to: {}".format(audio_temp_file))
# Add a text input for the title with a default value of 0
title = st.text_input("Key", value="0")
# Add a button to start the rendering process
if st.button("Render audio"):
  render_audio(ckpt_temp_file, config_temp_file, audio_temp_file, title)