1
0
mirror of synced 2024-11-28 01:10:56 +01:00
Retrieval-based-Voice-Conve.../infer/modules/uvr5/modules.py
github-actions[bot] 1696c5356b
chore(sync): merge dev into main (#1399)
* Update mdxnet.py

* Update modules.py

* Rename preprocess.py to vr.py

* Add files via upload

* deps: add av lib (#1391)

* Add files via upload

* chore(format): run black on dev (#1398)

Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>

---------

Co-authored-by: RVC-Boss <129054828+RVC-Boss@users.noreply.github.com>
Co-authored-by: Hiroto N <hironow365@gmail.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
2023-10-08 18:54:58 +08:00

109 lines
3.8 KiB
Python

import os
import traceback
import logging
logger = logging.getLogger(__name__)
import ffmpeg
import torch
from configs.config import Config
from infer.modules.uvr5.mdxnet import MDXNetDereverb
from infer.modules.uvr5.vr import AudioPre, AudioPreDeEcho
config = Config()
def uvr(model_name, inp_root, save_root_vocal, paths, save_root_ins, agg, format0):
infos = []
try:
inp_root = inp_root.strip(" ").strip('"').strip("\n").strip('"').strip(" ")
save_root_vocal = (
save_root_vocal.strip(" ").strip('"').strip("\n").strip('"').strip(" ")
)
save_root_ins = (
save_root_ins.strip(" ").strip('"').strip("\n").strip('"').strip(" ")
)
if model_name == "onnx_dereverb_By_FoxJoy":
pre_fun = MDXNetDereverb(15, config.device)
else:
func = AudioPre if "DeEcho" not in model_name else AudioPreDeEcho
pre_fun = func(
agg=int(agg),
model_path=os.path.join(
os.getenv("weight_uvr5_root"), model_name + ".pth"
),
device=config.device,
is_half=config.is_half,
)
is_hp3 = "HP3" in model_name
if inp_root != "":
paths = [os.path.join(inp_root, name) for name in os.listdir(inp_root)]
else:
paths = [path.name for path in paths]
for path in paths:
inp_path = os.path.join(inp_root, path)
need_reformat = 1
done = 0
try:
info = ffmpeg.probe(inp_path, cmd="ffprobe")
if (
info["streams"][0]["channels"] == 2
and info["streams"][0]["sample_rate"] == "44100"
):
need_reformat = 0
pre_fun._path_audio_(
inp_path, save_root_ins, save_root_vocal, format0, is_hp3=is_hp3
)
done = 1
except:
need_reformat = 1
traceback.print_exc()
if need_reformat == 1:
tmp_path = "%s/%s.reformatted.wav" % (
os.path.join(os.environ["TEMP"]),
os.path.basename(inp_path),
)
os.system(
"ffmpeg -i %s -vn -acodec pcm_s16le -ac 2 -ar 44100 %s -y"
% (inp_path, tmp_path)
)
inp_path = tmp_path
try:
if done == 0:
pre_fun._path_audio_(
inp_path, save_root_ins, save_root_vocal, format0
)
infos.append("%s->Success" % (os.path.basename(inp_path)))
yield "\n".join(infos)
except:
try:
if done == 0:
pre_fun._path_audio_(
inp_path, save_root_ins, save_root_vocal, format0
)
infos.append("%s->Success" % (os.path.basename(inp_path)))
yield "\n".join(infos)
except:
infos.append(
"%s->%s" % (os.path.basename(inp_path), traceback.format_exc())
)
yield "\n".join(infos)
except:
infos.append(traceback.format_exc())
yield "\n".join(infos)
finally:
try:
if model_name == "onnx_dereverb_By_FoxJoy":
del pre_fun.pred.model
del pre_fun.pred.model_
else:
del pre_fun.model
del pre_fun
except:
traceback.print_exc()
if torch.cuda.is_available():
torch.cuda.empty_cache()
logger.info("Executed torch.cuda.empty_cache()")
yield "\n".join(infos)