Add partitioning and improved resuming to the spleeter filtering
This commit is contained in:
parent
f2977d360c
commit
bb891a3a53
|
@ -78,20 +78,34 @@ def main():
|
|||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument('--path')
|
||||
parser.add_argument('--out')
|
||||
parser.add_argument('--resume', default=None)
|
||||
parser.add_argument('--partition_size', default=None)
|
||||
parser.add_argument('--partition', default=None)
|
||||
args = parser.parse_args()
|
||||
|
||||
src_dir = args.path
|
||||
out_file = args.out
|
||||
output_sample_rate=22050
|
||||
batch_size=16
|
||||
waiting_for_file = args.resume is not None
|
||||
resume_file = args.resume
|
||||
|
||||
audio_loader = AudioAdapter.default()
|
||||
files = find_audio_files(src_dir, include_nonwav=True)
|
||||
|
||||
# Partition files if needed.
|
||||
if args.partition_size is not None:
|
||||
psz = int(args.partition_size)
|
||||
prt = int(args.partition)
|
||||
files = files[prt*psz:(prt+1)*psz]
|
||||
|
||||
#separator = Separator('pretrained_models/2stems', input_sr=output_sample_rate)
|
||||
separator = Separator('spleeter:2stems')
|
||||
unacceptable_files = open(out_file, 'w')
|
||||
for path in tqdm(files):
|
||||
print(f"Processing {src_dir}")
|
||||
unacceptable_files = open(out_file, 'a')
|
||||
for e, path in enumerate(tqdm(files)):
|
||||
if waiting_for_file and resume_file not in path:
|
||||
continue
|
||||
waiting_for_file = False
|
||||
print(f"{e}: Processing {path}")
|
||||
spleeter_ld, sr = audio_loader.load(path, sample_rate=output_sample_rate)
|
||||
sep = separator.separate(spleeter_ld)
|
||||
vocals = sep['vocals']
|
||||
|
|
Loading…
Reference in New Issue
Block a user