This commit is contained in:
Ftps
2024-01-21 01:15:44 +09:00
parent c839cab165
commit 7e0326146d
7 changed files with 17 additions and 22 deletions

View File

@@ -6,11 +6,11 @@ import click
from rvc.cli.handler.infer import infer
from rvc.cli.handler.train import train
from rvc.cli.handler.uvr5 import uvr
from rvc.cli.utils.dlmodel import dlmodel
from rvc.cli.utils.env import env
from rvc.cli.utils.initialize import init
@click.group(
context_settings={"help_option_names": ["-h", "--help"]},
help="rvc cli feature list",
@@ -18,6 +18,7 @@ from rvc.cli.utils.initialize import init
def cli():
pass
def main():
cli.add_command(infer)
cli.add_command(train)

View File

@@ -7,7 +7,6 @@ from scipy.io import wavfile
from rvc.modules.vc.modules import VC
logging.getLogger("numba").setLevel(logging.WARNING)

View File

@@ -1,4 +1,5 @@
import urllib
import click

View File

@@ -6,9 +6,6 @@ download model and setup environmmnt file
import click
import click
@click.command()
def init():
pass

View File

@@ -18,15 +18,13 @@ try:
except Exception: # pylint: disable=broad-exception-caught
pass
import logging
from time import time as ttime
import torch.nn as nn
import torch.nn.functional as F
from librosa.filters import mel
from librosa.util import normalize, pad_center, tiny
from scipy.signal import get_window
from librosa.filters import mel
from time import time as ttime
logger = logging.getLogger(__name__)
@@ -161,7 +159,6 @@ class STFT(torch.nn.Module):
return reconstruction
class BiGRU(nn.Module):
def __init__(self, input_features, hidden_features, num_layers):
super(BiGRU, self).__init__()
@@ -415,8 +412,6 @@ class E2E(nn.Module):
return x
class MelSpectrogram(torch.nn.Module):
def __init__(
self,

View File

@@ -64,11 +64,11 @@ class Slicer:
def _apply_slice(self, waveform, begin, end):
if len(waveform.shape) > 1:
return waveform[
:, begin * self.hop_size: min(waveform.shape[1], end * self.hop_size)
:, begin * self.hop_size : min(waveform.shape[1], end * self.hop_size)
]
else:
return waveform[
begin * self.hop_size: min(waveform.shape[0], end * self.hop_size)
begin * self.hop_size : min(waveform.shape[0], end * self.hop_size)
]
# @timeit
@@ -106,7 +106,7 @@ class Slicer:
continue
# Need slicing. Record the range of silent frames to be removed.
if i - silence_start <= self.max_sil_kept:
pos = rms_list[silence_start: i + 1].argmin() + silence_start
pos = rms_list[silence_start : i + 1].argmin() + silence_start
if silence_start == 0:
sil_tags.append((0, pos))
else:
@@ -114,17 +114,17 @@ class Slicer:
clip_start = pos
elif i - silence_start <= self.max_sil_kept * 2:
pos = rms_list[
i - self.max_sil_kept: silence_start + self.max_sil_kept + 1
i - self.max_sil_kept : silence_start + self.max_sil_kept + 1
].argmin()
pos += i - self.max_sil_kept
pos_l = (
rms_list[
silence_start: silence_start + self.max_sil_kept + 1
silence_start : silence_start + self.max_sil_kept + 1
].argmin()
+ silence_start
)
pos_r = (
rms_list[i - self.max_sil_kept: i + 1].argmin()
rms_list[i - self.max_sil_kept : i + 1].argmin()
+ i
- self.max_sil_kept
)
@@ -137,12 +137,12 @@ class Slicer:
else:
pos_l = (
rms_list[
silence_start: silence_start + self.max_sil_kept + 1
silence_start : silence_start + self.max_sil_kept + 1
].argmin()
+ silence_start
)
pos_r = (
rms_list[i - self.max_sil_kept: i + 1].argmin()
rms_list[i - self.max_sil_kept : i + 1].argmin()
+ i
- self.max_sil_kept
)
@@ -159,7 +159,7 @@ class Slicer:
and total_frames - silence_start >= self.min_interval
):
silence_end = min(total_frames, silence_start + self.max_sil_kept)
pos = rms_list[silence_start: silence_end + 1].argmin() + silence_start
pos = rms_list[silence_start : silence_end + 1].argmin() + silence_start
sil_tags.append((pos, total_frames + 1))
# Apply and return slices.
if len(sil_tags) == 0:

View File

@@ -9,7 +9,9 @@ def export_onnx(ModelPath, ExportedPath):
vec_channels = 256 if cpt.get("version", "v1") == "v1" else 768
test_phone = torch.rand(1, 200, vec_channels) # hidden unit
test_phone_lengths = torch.tensor([200]).long() # hidden unit length (doesn't seem to help)
test_phone_lengths = torch.tensor(
[200]
).long() # hidden unit length (doesn't seem to help)
test_pitch = torch.randint(size=(1, 200), low=5, high=255) # Base frequency (in Hz)
test_pitchf = torch.rand(1, 200) # nsf base frequency
test_ds = torch.LongTensor([0]) # Speaker ID