fumiama/Retrieval-based-Voice-Conversion-WebUI
1
0
Fork 0
mirror of https://github.com/fumiama/Retrieval-based-Voice-Conversion-WebUI.git synced 2026-06-08 03:55:47 +08:00
Code Activity

optimize(uvr5): apply jit to spec_utils & fix flac save

Browse Source 
also fix #85
This commit is contained in:
源文雨
2024-11-28 23:19:05 +09:00
parent 4582d4b49a
commit 5969314e8d
11 changed files with 104 additions and 581 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
infer/modules/uvr5/mdxnet.py
View File

@@ -3,12 +3,11 @@ import logging
logger = logging.getLogger(__name__)
import librosa
import numpy as np
import torch
from tqdm import tqdm
from infer.lib.audio import downsample_audio, save_audio
from infer.lib.audio import load_audio, save_audio
cpu = torch.device("cpu")
@@ -201,29 +200,18 @@ class Predictor:
os.makedirs(vocal_root, exist_ok=True)
os.makedirs(others_root, exist_ok=True)
basename = os.path.basename(m)
mix, rate = librosa.load(m, mono=False, sr=44100)
mix, rate = load_audio(m, mono=False, sr=44100)
if mix.ndim == 1:
mix = np.asfortranarray([mix, mix])
mix = mix.T
sources = self.demix(mix.T)
opt = sources[0].T
if format in ["wav", "flac"]:
save_audio(
"%s/vocal_%s.%s" % (vocal_root, basename, format), mix - opt, rate
)
save_audio(
"%s/instrument_%s.%s" % (others_root, basename, format), opt, rate
)
else:
path_vocal = "%s/vocal_%s.wav" % (vocal_root, basename)
path_other = "%s/instrument_%s.wav" % (others_root, basename)
save_audio(path_vocal, opt, rate)
save_audio(path_other, opt, rate)
opt_path_vocal = path_vocal[:-4] + ".%s" % format
opt_path_other = path_other[:-4] + ".%s" % format
downsample_audio(path_vocal, opt_path_vocal, format)
downsample_audio(path_other, opt_path_other, format)
save_audio(
"%s/vocal_%s.%s" % (vocal_root, basename, format), mix - opt, rate, True, format=format,
)
save_audio(
"%s/instrument_%s.%s" % (others_root, basename, format), opt, rate, True, format=format,
)
class MDXNetDereverb:
def __init__(self, chunks, device):