updated process_datasets.py, added argparsing so I can mostly stop manually editing things, and some other cleanup
This commit is contained in:
parent
debcc93e7e
commit
7cdfa3dc0c
|
|
@ -1,191 +1,104 @@
|
||||||
import os
|
import os
|
||||||
import json
|
import json
|
||||||
|
import argparse
|
||||||
import torch
|
import torch
|
||||||
import torchaudio
|
import torchaudio
|
||||||
import numpy as np
|
import numpy as np
|
||||||
|
|
||||||
from tqdm.auto import tqdm
|
from tqdm.auto import tqdm
|
||||||
from pathlib import Path
|
from pathlib import Path
|
||||||
from vall_e.config import cfg
|
from vall_e.config import cfg
|
||||||
|
|
||||||
# things that could be args
|
|
||||||
cfg.sample_rate = 24_000
|
|
||||||
cfg.audio_backend = "encodec"
|
|
||||||
"""
|
|
||||||
cfg.inference.weight_dtype = "bfloat16"
|
|
||||||
cfg.inference.dtype = torch.bfloat16
|
|
||||||
cfg.inference.amp = True
|
|
||||||
"""
|
|
||||||
|
|
||||||
from vall_e.emb.g2p import encode as valle_phonemize
|
|
||||||
from vall_e.emb.qnt import encode as valle_quantize, _replace_file_extension
|
|
||||||
|
|
||||||
input_audio = "voices"
|
|
||||||
input_metadata = "metadata"
|
|
||||||
output_dataset = f"training-{'2' if cfg.sample_rate == 24_000 else '4'}{'8' if cfg.sample_rate == 48_000 else '4'}KHz-{cfg.audio_backend}"
|
|
||||||
device = "cuda"
|
|
||||||
|
|
||||||
audio_extension = ".enc"
|
|
||||||
if cfg.audio_backend == "dac":
|
|
||||||
audio_extension = ".dac"
|
|
||||||
elif cfg.audio_backend == "audiodec":
|
|
||||||
audio_extension = ".dec"
|
|
||||||
|
|
||||||
slice = "auto"
|
|
||||||
missing = {
|
|
||||||
"transcription": [],
|
|
||||||
"audio": []
|
|
||||||
}
|
|
||||||
dataset = []
|
|
||||||
|
|
||||||
def pad(num, zeroes):
|
def pad(num, zeroes):
|
||||||
return str(num).zfill(zeroes+1)
|
return str(num).zfill(zeroes+1)
|
||||||
|
|
||||||
for dataset_name in sorted(os.listdir(f'./{input_audio}/')):
|
def process_items( items, stride=0 ):
|
||||||
if not os.path.isdir(f'./{input_audio}/{dataset_name}/'):
|
items = sorted( items )
|
||||||
print("Is not dir:", f'./{input_audio}/{dataset_name}/')
|
return items if stride == 0 else [ item for i, item in enumerate( items ) if i % stride == 0 ]
|
||||||
continue
|
|
||||||
|
|
||||||
for speaker_id in tqdm(sorted(os.listdir(f'./{input_audio}/{dataset_name}/')), desc=f"Processing speaker in {dataset_name}"):
|
def process_dataset( args ):
|
||||||
if not os.path.isdir(f'./{input_audio}/{dataset_name}/{speaker_id}'):
|
# encodec / vocos
|
||||||
print("Is not dir:", f'./{input_audio}/{dataset_name}/{speaker_id}')
|
|
||||||
continue
|
|
||||||
|
|
||||||
os.makedirs(f'./{output_dataset}/{dataset_name}/{speaker_id}/', exist_ok=True)
|
|
||||||
|
|
||||||
if speaker_id == "Noise":
|
if args.audio_backend in ["encodec", "vocos"]:
|
||||||
for filename in sorted(os.listdir(f'./{input_audio}/{dataset_name}/{speaker_id}/')):
|
audio_extension = ".enc"
|
||||||
inpath = Path(f'./{input_audio}/{dataset_name}/{speaker_id}/{filename}')
|
cfg.sample_rate = 24_000
|
||||||
outpath = Path(f'./{output_dataset}/{dataset_name}/{speaker_id}/{filename}')
|
cfg.model.resp_levels = 8
|
||||||
|
elif args.audio_backend == "dac":
|
||||||
|
audio_extension = ".dac"
|
||||||
|
cfg.sample_rate = 44_100
|
||||||
|
cfg.model.resp_levels = 9
|
||||||
|
elif cfg.audio_backend == "audiodec":
|
||||||
|
sample_rate = 48_000
|
||||||
|
audio_extension = ".dec"
|
||||||
|
cfg.model.resp_levels = 8 # ?
|
||||||
|
else:
|
||||||
|
raise Exception(f"Unknown audio backend: {args.audio_backend}")
|
||||||
|
|
||||||
if _replace_file_extension(outpath, audio_extension).exists():
|
# prepare from args
|
||||||
continue
|
cfg.audio_backend = args.audio_backend # "encodec"
|
||||||
|
cfg.inference.weight_dtype = args.dtype # "bfloat16"
|
||||||
|
cfg.inference.amp = args.amp # False
|
||||||
|
|
||||||
waveform, sample_rate = torchaudio.load(inpath)
|
# import after because we've overriden the config above
|
||||||
qnt = valle_quantize(waveform, sr=sample_rate, device=device)
|
from vall_e.emb.g2p import encode as valle_phonemize
|
||||||
|
from vall_e.emb.qnt import encode as valle_quantize, _replace_file_extension
|
||||||
|
|
||||||
if cfg.audio_backend == "dac":
|
input_audio = args.input_audio # "voice""
|
||||||
np.save(open(_replace_file_extension(outpath, audio_extension), "wb"), {
|
input_metadata = args.input_metadata # "metadata"
|
||||||
"codes": qnt.codes.cpu().numpy().astype(np.uint16),
|
output_group = f"{args.output_group}-{'2' if cfg.sample_rate == 24_000 else '4'}{'8' if cfg.sample_rate == 48_000 else '4'}KHz-{cfg.audio_backend}" # "training"
|
||||||
"metadata": {
|
device = args.device # "cuda"
|
||||||
"original_length": qnt.original_length,
|
raise_exceptions = args.raise_exceptions # False
|
||||||
"sample_rate": qnt.sample_rate,
|
stride = args.stride # 0
|
||||||
|
slice = args.slice # "auto"
|
||||||
"input_db": qnt.input_db.cpu().numpy().astype(np.float32),
|
|
||||||
"chunk_length": qnt.chunk_length,
|
|
||||||
"channels": qnt.channels,
|
|
||||||
"padding": qnt.padding,
|
|
||||||
"dac_version": "1.0.0",
|
|
||||||
},
|
|
||||||
})
|
|
||||||
else:
|
|
||||||
np.save(open(_replace_file_extension(outpath, audio_extension), "wb"), {
|
|
||||||
"codes": qnt.cpu().numpy().astype(np.uint16),
|
|
||||||
"metadata": {
|
|
||||||
"original_length": waveform.shape[-1],
|
|
||||||
"sample_rate": sample_rate,
|
|
||||||
},
|
|
||||||
})
|
|
||||||
|
|
||||||
continue
|
language_map = {} # k = group, v = language
|
||||||
|
|
||||||
metadata_path = Path(f'./{input_metadata}/{dataset_name}/{speaker_id}/whisper.json')
|
ignore_groups = [] # skip these groups
|
||||||
if not metadata_path.exists():
|
ignore_speakers = [] # skip these speakers
|
||||||
missing["transcription"].append(str(metadata_path))
|
|
||||||
|
only_groups = [] # only process these groups
|
||||||
|
only_speakers = [] # only process these speakers
|
||||||
|
|
||||||
|
always_slice_groups = [] # always slice from this group
|
||||||
|
|
||||||
|
missing = {
|
||||||
|
"transcription": [],
|
||||||
|
"audio": []
|
||||||
|
}
|
||||||
|
dataset = []
|
||||||
|
|
||||||
|
for group_name in sorted(os.listdir(f'./{input_audio}/')):
|
||||||
|
if not os.path.isdir(f'./{input_audio}/{group_name}/'):
|
||||||
|
print("Is not dir:", f'./{input_audio}/{group_name}/')
|
||||||
continue
|
continue
|
||||||
|
|
||||||
try:
|
if group_name in ignore_groups:
|
||||||
metadata = json.loads(open(metadata_path, "r", encoding="utf-8").read())
|
continue
|
||||||
except Exception as e:
|
if only_groups and group_name not in only_groups:
|
||||||
missing["transcription"].append(str(metadata_path))
|
|
||||||
continue
|
continue
|
||||||
|
|
||||||
if f'{dataset_name}/{speaker_id}' not in dataset:
|
for speaker_id in tqdm(process_items(os.listdir(f'./{input_audio}/{group_name}/'), stride=stride), desc=f"Processing speaker in {group_name}"):
|
||||||
dataset.append(f'{dataset_name}/{speaker_id}')
|
if not os.path.isdir(f'./{input_audio}/{group_name}/{speaker_id}'):
|
||||||
|
print("Is not dir:", f'./{input_audio}/{group_name}/{speaker_id}')
|
||||||
txts = []
|
|
||||||
wavs = []
|
|
||||||
|
|
||||||
use_slices = slice == True or (slice == "auto" and len(metadata.keys()) == 1) or dataset_name in ["LibriVox", "Audiobooks"]
|
|
||||||
|
|
||||||
for filename in sorted(metadata.keys()):
|
|
||||||
inpath = Path(f'./{input_audio}/{dataset_name}/{speaker_id}/{filename}')
|
|
||||||
if not inpath.exists():
|
|
||||||
missing["audio"].append(str(inpath))
|
|
||||||
continue
|
continue
|
||||||
|
|
||||||
extension = os.path.splitext(filename)[-1][1:]
|
if speaker_id in ignore_speakers:
|
||||||
fname = filename.replace(f'.{extension}', "")
|
continue
|
||||||
|
if only_speakers and speaker_id not in only_speakers:
|
||||||
|
continue
|
||||||
|
|
||||||
waveform, sample_rate = None, None
|
os.makedirs(f'./{output_group}/{group_name}/{speaker_id}/', exist_ok=True)
|
||||||
language = metadata[filename]["language"] if "language" in metadata[filename] else "en"
|
|
||||||
|
|
||||||
if len(metadata[filename]["segments"]) == 0 or not use_slices:
|
if speaker_id == "Noise":
|
||||||
outpath = Path(f'./{output_dataset}/{dataset_name}/{speaker_id}/{fname}.{extension}')
|
for filename in sorted(os.listdir(f'./{input_audio}/{group_name}/{speaker_id}/')):
|
||||||
text = metadata[filename]["text"]
|
inpath = Path(f'./{input_audio}/{group_name}/{speaker_id}/{filename}')
|
||||||
|
outpath = Path(f'./{output_group}/{group_name}/{speaker_id}/{filename}')
|
||||||
if len(text) == 0:
|
|
||||||
continue
|
|
||||||
|
|
||||||
if _replace_file_extension(outpath, audio_extension).exists():
|
|
||||||
continue
|
|
||||||
|
|
||||||
if waveform is None:
|
|
||||||
waveform, sample_rate = torchaudio.load(inpath)
|
|
||||||
if waveform.shape[0] > 1:
|
|
||||||
waveform = torch.mean(waveform, dim=0, keepdim=True)
|
|
||||||
|
|
||||||
wavs.append((
|
|
||||||
outpath,
|
|
||||||
text,
|
|
||||||
language,
|
|
||||||
waveform,
|
|
||||||
sample_rate
|
|
||||||
))
|
|
||||||
else:
|
|
||||||
i = 0
|
|
||||||
for segment in metadata[filename]["segments"]:
|
|
||||||
id = pad(i, 4)
|
|
||||||
i = i + 1
|
|
||||||
|
|
||||||
outpath = Path(f'./{output_dataset}/{dataset_name}/{speaker_id}/{fname}_{id}.{extension}')
|
|
||||||
text = segment["text"]
|
|
||||||
|
|
||||||
if len(text) == 0:
|
|
||||||
continue
|
|
||||||
|
|
||||||
if _replace_file_extension(outpath, audio_extension).exists():
|
if _replace_file_extension(outpath, audio_extension).exists():
|
||||||
continue
|
continue
|
||||||
|
|
||||||
if waveform is None:
|
waveform, sample_rate = torchaudio.load(inpath)
|
||||||
waveform, sample_rate = torchaudio.load(inpath)
|
|
||||||
if waveform.shape[0] > 1:
|
|
||||||
waveform = torch.mean(waveform, dim=0, keepdim=True)
|
|
||||||
|
|
||||||
start = int(segment['start'] * sample_rate)
|
|
||||||
end = int(segment['end'] * sample_rate)
|
|
||||||
|
|
||||||
if start < 0:
|
|
||||||
start = 0
|
|
||||||
if end >= waveform.shape[-1]:
|
|
||||||
end = waveform.shape[-1] - 1
|
|
||||||
|
|
||||||
if end - start < 0:
|
|
||||||
continue
|
|
||||||
|
|
||||||
wavs.append((
|
|
||||||
outpath,
|
|
||||||
text,
|
|
||||||
language,
|
|
||||||
waveform[:, start:end],
|
|
||||||
sample_rate
|
|
||||||
))
|
|
||||||
|
|
||||||
if len(wavs) > 0:
|
|
||||||
for job in tqdm(wavs, desc=f"Quantizing: {speaker_id}"):
|
|
||||||
try:
|
|
||||||
outpath, text, language, waveform, sample_rate = job
|
|
||||||
|
|
||||||
phones = valle_phonemize( text, language=language )
|
|
||||||
qnt = valle_quantize(waveform, sr=sample_rate, device=device)
|
qnt = valle_quantize(waveform, sr=sample_rate, device=device)
|
||||||
|
|
||||||
if cfg.audio_backend == "dac":
|
if cfg.audio_backend == "dac":
|
||||||
|
|
@ -200,10 +113,6 @@ for dataset_name in sorted(os.listdir(f'./{input_audio}/')):
|
||||||
"channels": qnt.channels,
|
"channels": qnt.channels,
|
||||||
"padding": qnt.padding,
|
"padding": qnt.padding,
|
||||||
"dac_version": "1.0.0",
|
"dac_version": "1.0.0",
|
||||||
|
|
||||||
"text": text.strip(),
|
|
||||||
"phonemes": "".join(phones),
|
|
||||||
"language": language,
|
|
||||||
},
|
},
|
||||||
})
|
})
|
||||||
else:
|
else:
|
||||||
|
|
@ -212,15 +121,168 @@ for dataset_name in sorted(os.listdir(f'./{input_audio}/')):
|
||||||
"metadata": {
|
"metadata": {
|
||||||
"original_length": waveform.shape[-1],
|
"original_length": waveform.shape[-1],
|
||||||
"sample_rate": sample_rate,
|
"sample_rate": sample_rate,
|
||||||
|
|
||||||
"text": text.strip(),
|
|
||||||
"phonemes": "".join(phones),
|
|
||||||
"language": language,
|
|
||||||
},
|
},
|
||||||
})
|
})
|
||||||
except Exception as e:
|
|
||||||
print(f"Failed to quantize: {outpath}:", e)
|
|
||||||
continue
|
|
||||||
|
|
||||||
open("./missing.json", 'w', encoding='utf-8').write(json.dumps(missing))
|
continue
|
||||||
open("./dataset_list.json", 'w', encoding='utf-8').write(json.dumps(dataset))
|
|
||||||
|
metadata_path = Path(f'./{input_metadata}/{group_name}/{speaker_id}/whisper.json')
|
||||||
|
if not metadata_path.exists():
|
||||||
|
missing["transcription"].append(str(metadata_path))
|
||||||
|
continue
|
||||||
|
|
||||||
|
try:
|
||||||
|
metadata = json.loads(open(metadata_path, "r", encoding="utf-8").read())
|
||||||
|
except Exception as e:
|
||||||
|
missing["transcription"].append(str(metadata_path))
|
||||||
|
continue
|
||||||
|
|
||||||
|
if f'{group_name}/{speaker_id}' not in dataset:
|
||||||
|
dataset.append(f'{group_name}/{speaker_id}')
|
||||||
|
|
||||||
|
txts = []
|
||||||
|
wavs = []
|
||||||
|
|
||||||
|
use_slices = slice == True or (slice == "auto" and len(metadata.keys()) == 1) or group_name in always_slice_groups
|
||||||
|
|
||||||
|
for filename in sorted(metadata.keys()):
|
||||||
|
inpath = Path(f'./{input_audio}/{group_name}/{speaker_id}/{filename}')
|
||||||
|
if not inpath.exists():
|
||||||
|
missing["audio"].append(str(inpath))
|
||||||
|
continue
|
||||||
|
|
||||||
|
extension = os.path.splitext(filename)[-1][1:]
|
||||||
|
fname = filename.replace(f'.{extension}', "")
|
||||||
|
|
||||||
|
waveform, sample_rate = None, None
|
||||||
|
language = language_map[group_name] if group_name in language_map else (metadata[filename]["language"] if "language" in metadata[filename] else "en")
|
||||||
|
|
||||||
|
if len(metadata[filename]["segments"]) == 0 or not use_slices:
|
||||||
|
outpath = Path(f'./{output_group}/{group_name}/{speaker_id}/{fname}.{extension}')
|
||||||
|
text = metadata[filename]["text"]
|
||||||
|
|
||||||
|
if len(text) == 0:
|
||||||
|
continue
|
||||||
|
|
||||||
|
if _replace_file_extension(outpath, audio_extension).exists():
|
||||||
|
continue
|
||||||
|
|
||||||
|
if waveform is None:
|
||||||
|
waveform, sample_rate = torchaudio.load(inpath)
|
||||||
|
if waveform.shape[0] > 1:
|
||||||
|
waveform = torch.mean(waveform, dim=0, keepdim=True)
|
||||||
|
|
||||||
|
wavs.append((
|
||||||
|
outpath,
|
||||||
|
text,
|
||||||
|
language,
|
||||||
|
waveform,
|
||||||
|
sample_rate
|
||||||
|
))
|
||||||
|
else:
|
||||||
|
i = 0
|
||||||
|
for segment in metadata[filename]["segments"]:
|
||||||
|
id = pad(i, 4)
|
||||||
|
i = i + 1
|
||||||
|
|
||||||
|
outpath = Path(f'./{output_group}/{group_name}/{speaker_id}/{fname}_{id}.{extension}')
|
||||||
|
text = segment["text"]
|
||||||
|
|
||||||
|
if len(text) == 0:
|
||||||
|
continue
|
||||||
|
|
||||||
|
if _replace_file_extension(outpath, audio_extension).exists():
|
||||||
|
continue
|
||||||
|
|
||||||
|
if waveform is None:
|
||||||
|
waveform, sample_rate = torchaudio.load(inpath)
|
||||||
|
if waveform.shape[0] > 1:
|
||||||
|
waveform = torch.mean(waveform, dim=0, keepdim=True)
|
||||||
|
|
||||||
|
start = int(segment['start'] * sample_rate)
|
||||||
|
end = int(segment['end'] * sample_rate)
|
||||||
|
|
||||||
|
if start < 0:
|
||||||
|
start = 0
|
||||||
|
if end >= waveform.shape[-1]:
|
||||||
|
end = waveform.shape[-1] - 1
|
||||||
|
|
||||||
|
if end - start < 0:
|
||||||
|
continue
|
||||||
|
|
||||||
|
wavs.append((
|
||||||
|
outpath,
|
||||||
|
text,
|
||||||
|
language,
|
||||||
|
waveform[:, start:end],
|
||||||
|
sample_rate
|
||||||
|
))
|
||||||
|
|
||||||
|
if len(wavs) > 0:
|
||||||
|
for job in tqdm(wavs, desc=f"Quantizing: {speaker_id}"):
|
||||||
|
try:
|
||||||
|
outpath, text, language, waveform, sample_rate = job
|
||||||
|
|
||||||
|
phones = valle_phonemize(text, language=language)
|
||||||
|
qnt = valle_quantize(waveform, sr=sample_rate, device=device)
|
||||||
|
|
||||||
|
|
||||||
|
if cfg.audio_backend == "dac":
|
||||||
|
np.save(open(_replace_file_extension(outpath, audio_extension), "wb"), {
|
||||||
|
"codes": qnt.codes.cpu().numpy().astype(np.uint16),
|
||||||
|
"metadata": {
|
||||||
|
"original_length": qnt.original_length,
|
||||||
|
"sample_rate": qnt.sample_rate,
|
||||||
|
|
||||||
|
"input_db": qnt.input_db.cpu().numpy().astype(np.float32),
|
||||||
|
"chunk_length": qnt.chunk_length,
|
||||||
|
"channels": qnt.channels,
|
||||||
|
"padding": qnt.padding,
|
||||||
|
"dac_version": "1.0.0",
|
||||||
|
|
||||||
|
"text": text.strip(),
|
||||||
|
"phonemes": "".join(phones),
|
||||||
|
"language": language,
|
||||||
|
},
|
||||||
|
})
|
||||||
|
else:
|
||||||
|
np.save(open(_replace_file_extension(outpath, audio_extension), "wb"), {
|
||||||
|
"codes": qnt.cpu().numpy().astype(np.uint16),
|
||||||
|
"metadata": {
|
||||||
|
"original_length": waveform.shape[-1],
|
||||||
|
"sample_rate": sample_rate,
|
||||||
|
|
||||||
|
"text": text.strip(),
|
||||||
|
"phonemes": "".join(phones),
|
||||||
|
"language": language,
|
||||||
|
},
|
||||||
|
})
|
||||||
|
except Exception as e:
|
||||||
|
print(f"Failed to quantize: {outpath}:", e)
|
||||||
|
if raise_exceptions:
|
||||||
|
raise e
|
||||||
|
continue
|
||||||
|
|
||||||
|
open("./missing.json", 'w', encoding='utf-8').write(json.dumps(missing))
|
||||||
|
open("./dataset_list.json", 'w', encoding='utf-8').write(json.dumps(dataset))
|
||||||
|
|
||||||
|
def main():
|
||||||
|
parser = argparse.ArgumentParser()
|
||||||
|
|
||||||
|
parser.add_argument("--audio-backend", type=str, default="encodec")
|
||||||
|
parser.add_argument("--dtype", type=str, default="bfloat16")
|
||||||
|
parser.add_argument("--amp", action="store_true")
|
||||||
|
parser.add_argument("--input-audio", type=str, default="voices")
|
||||||
|
parser.add_argument("--input-metadata", type=str, default="metadata")
|
||||||
|
parser.add_argument("--output_group", type=str, default="training")
|
||||||
|
parser.add_argument("--device", type=str, default="cuda")
|
||||||
|
parser.add_argument("--raise-exceptions", action="store_true")
|
||||||
|
parser.add_argument("--stride", type=int, default=0)
|
||||||
|
parser.add_argument("--slice", type=str, default="auto")
|
||||||
|
|
||||||
|
args = parser.parse_args()
|
||||||
|
|
||||||
|
process_dataset( args )
|
||||||
|
|
||||||
|
if __name__ == "__main__":
|
||||||
|
main()
|
||||||
|
|
@ -813,7 +813,7 @@ class Base(nn.Module):
|
||||||
inputs_embeds=x,
|
inputs_embeds=x,
|
||||||
past_key_values=state,
|
past_key_values=state,
|
||||||
position_ids=position_ids,
|
position_ids=position_ids,
|
||||||
use_cache=True,
|
use_cache=not self.training,
|
||||||
# return_dict=True,
|
# return_dict=True,
|
||||||
)
|
)
|
||||||
if self.n_experts > 1 and self.training:
|
if self.n_experts > 1 and self.training:
|
||||||
|
|
@ -1350,15 +1350,6 @@ class Base(nn.Module):
|
||||||
x, m = list_to_tensor(x_list)
|
x, m = list_to_tensor(x_list)
|
||||||
|
|
||||||
training = self.training
|
training = self.training
|
||||||
# yes, there's a better way.
|
|
||||||
"""
|
|
||||||
training = False
|
|
||||||
for batch_index, batch in enumerate(inputs):
|
|
||||||
for name, input in batch:
|
|
||||||
if name == "targ":
|
|
||||||
training = True
|
|
||||||
"""
|
|
||||||
|
|
||||||
device = x.device
|
device = x.device
|
||||||
batch_size = len(x_list)
|
batch_size = len(x_list)
|
||||||
|
|
||||||
|
|
|
||||||
Loading…
Reference in New Issue
Block a user