def main(args): # pylint: disable=redefined-outer-name
# pylint: disable=global-variable-undefined
global meta_data_train, meta_data_eval, speaker_mapping, symbols, phonemes, model_characters
# Audio processor
ap = AudioProcessor(**c.audio)
# setup custom characters if set in config file.
if "characters" in c.keys():
symbols, phonemes = make_symbols(**c.characters)
# DISTRUBUTED
if num_gpus > 1:
init_distributed(args.rank, num_gpus, args.group_id,
c.distributed["backend"], c.distributed["url"])
num_chars = len(phonemes) if c.use_phonemes else len(symbols)
model_characters = phonemes if c.use_phonemes else symbols
# load data instances
meta_data_train, meta_data_eval = load_meta_data(c.datasets)
# set the portion of the data used for training
if "train_portion" in c.keys():
meta_data_train = meta_data_train[:int(
len(meta_data_train) * c.train_portion)]
if "eval_portion" in c.keys():
meta_data_eval = meta_data_eval[:int(
len(meta_data_eval) * c.eval_portion)]
# parse speakers
num_speakers, speaker_embedding_dim, speaker_mapping = parse_speakers(
c, args, meta_data_train, OUT_PATH)
model = setup_model(num_chars, num_speakers, c, speaker_embedding_dim)
# scalers for mixed precision training
scaler = torch.cuda.amp.GradScaler() if c.mixed_precision else None
scaler_st = torch.cuda.amp.GradScaler(
) if c.mixed_precision and c.separate_stopnet else None
params = set_weight_decay(model, c.wd)
optimizer = RAdam(params, lr=c.lr, weight_decay=0)
if c.stopnet and c.separate_stopnet:
optimizer_st = RAdam(model.decoder.stopnet.parameters(),
lr=c.lr,
weight_decay=0)
else:
optimizer_st = None
# setup criterion
criterion = TacotronLoss(c,
stopnet_pos_weight=c.stopnet_pos_weight,
ga_sigma=0.4)
if args.restore_path:
print(f" > Restoring from {os.path.basename(args.restore_path)}...")
checkpoint = torch.load(args.restore_path, map_location="cpu")
try:
print(" > Restoring Model...")
model.load_state_dict(checkpoint["model"])
# optimizer restore
print(" > Restoring Optimizer...")
optimizer.load_state_dict(checkpoint["optimizer"])
if "scaler" in checkpoint and c.mixed_precision:
print(" > Restoring AMP Scaler...")
scaler.load_state_dict(checkpoint["scaler"])
if c.reinit_layers:
raise RuntimeError
except (KeyError, RuntimeError):
print(" > Partial model initialization...")
model_dict = model.state_dict()
model_dict = set_init_dict(model_dict, checkpoint["model"], c)
# torch.save(model_dict, os.path.join(OUT_PATH, 'state_dict.pt'))
# print("State Dict saved for debug in: ", os.path.join(OUT_PATH, 'state_dict.pt'))
model.load_state_dict(model_dict)
del model_dict
for group in optimizer.param_groups:
group["lr"] = c.lr
print(" > Model restored from step %d" % checkpoint["step"],
flush=True)
args.restore_step = checkpoint["step"]
else:
args.restore_step = 0
if use_cuda:
model.cuda()
criterion.cuda()
# DISTRUBUTED
if num_gpus > 1:
model = apply_gradient_allreduce(model)
if c.noam_schedule:
scheduler = NoamLR(optimizer,
warmup_steps=c.warmup_steps,
last_epoch=args.restore_step - 1)
else:
scheduler = None
num_params = count_parameters(model)
print("\n > Model has {} parameters".format(num_params), flush=True)
if args.restore_step == 0 or not args.best_path:
best_loss = float("inf")
print(" > Starting with inf best loss.")
else:
print(" > Restoring best loss from "
f"{os.path.basename(args.best_path)} ...")
best_loss = torch.load(args.best_path,
map_location="cpu")["model_loss"]
print(f" > Starting with loaded last best loss {best_loss}.")
keep_all_best = c.get("keep_all_best", False)
keep_after = c.get("keep_after", 10000) # void if keep_all_best False
# define data loaders
train_loader = setup_loader(ap,
model.decoder.r,
is_val=False,
verbose=True)
eval_loader = setup_loader(ap, model.decoder.r, is_val=True)
global_step = args.restore_step
for epoch in range(0, c.epochs):
c_logger.print_epoch_start(epoch, c.epochs)
# set gradual training
if c.gradual_training is not None:
r, c.batch_size = gradual_training_scheduler(global_step, c)
c.r = r
model.decoder.set_r(r)
if c.bidirectional_decoder:
model.decoder_backward.set_r(r)
train_loader.dataset.outputs_per_step = r
eval_loader.dataset.outputs_per_step = r
train_loader = setup_loader(ap,
model.decoder.r,
is_val=False,
dataset=train_loader.dataset)
eval_loader = setup_loader(ap,
model.decoder.r,
is_val=True,
dataset=eval_loader.dataset)
print("\n > Number of output frames:", model.decoder.r)
# train one epoch
train_avg_loss_dict, global_step = train(
train_loader,
model,
criterion,
optimizer,
optimizer_st,
scheduler,
ap,
global_step,
epoch,
scaler,
scaler_st,
)
# eval one epoch
eval_avg_loss_dict = evaluate(eval_loader, model, criterion, ap,
global_step, epoch)
c_logger.print_epoch_end(epoch, eval_avg_loss_dict)
target_loss = train_avg_loss_dict["avg_postnet_loss"]
if c.run_eval:
target_loss = eval_avg_loss_dict["avg_postnet_loss"]
best_loss = save_best_model(
target_loss,
best_loss,
model,
optimizer,
global_step,
epoch,
c.r,
OUT_PATH,
model_characters,
keep_all_best=keep_all_best,
keep_after=keep_after,
scaler=scaler.state_dict() if c.mixed_precision else None,
)
def get_criterion(self) -> nn.Module:
"""Get the model criterion used in training."""
return TacotronLoss(self.config)
def main(args): # pylint: disable=redefined-outer-name
# pylint: disable=global-variable-undefined
global meta_data_train, meta_data_eval, symbols, phonemes
# Audio processor
ap = AudioProcessor(**c.audio)
if 'characters' in c.keys():
symbols, phonemes = make_symbols(**c.characters)
# DISTRUBUTED
if num_gpus > 1:
init_distributed(args.rank, num_gpus, args.group_id,
c.distributed["backend"], c.distributed["url"])
num_chars = len(phonemes) if c.use_phonemes else len(symbols)
# load data instances
meta_data_train, meta_data_eval = load_meta_data(c.datasets)
# set the portion of the data used for training
if 'train_portion' in c.keys():
meta_data_train = meta_data_train[:int(
len(meta_data_train) * c.train_portion)]
if 'eval_portion' in c.keys():
meta_data_eval = meta_data_eval[:int(
len(meta_data_eval) * c.eval_portion)]
# parse speakers
if c.use_speaker_embedding:
speakers = get_speakers(meta_data_train)
if args.restore_path:
if c.use_external_speaker_embedding_file: # if restore checkpoint and use External Embedding file
prev_out_path = os.path.dirname(args.restore_path)
speaker_mapping = load_speaker_mapping(prev_out_path)
if not speaker_mapping:
print(
"WARNING: speakers.json was not found in restore_path, trying to use CONFIG.external_speaker_embedding_file"
)
speaker_mapping = load_speaker_mapping(
c.external_speaker_embedding_file)
if not speaker_mapping:
raise RuntimeError(
"You must copy the file speakers.json to restore_path, or set a valid file in CONFIG.external_speaker_embedding_file"
)
speaker_embedding_dim = len(speaker_mapping[list(
speaker_mapping.keys())[0]]['embedding'])
elif not c.use_external_speaker_embedding_file: # if restore checkpoint and don't use External Embedding file
prev_out_path = os.path.dirname(args.restore_path)
speaker_mapping = load_speaker_mapping(prev_out_path)
speaker_embedding_dim = None
assert all([speaker in speaker_mapping
for speaker in speakers]), "As of now you, you cannot " \
"introduce new speakers to " \
"a previously trained model."
elif c.use_external_speaker_embedding_file and c.external_speaker_embedding_file: # if start new train using External Embedding file
speaker_mapping = load_speaker_mapping(
c.external_speaker_embedding_file)
speaker_embedding_dim = len(speaker_mapping[list(
speaker_mapping.keys())[0]]['embedding'])
elif c.use_external_speaker_embedding_file and not c.external_speaker_embedding_file: # if start new train using External Embedding file and don't pass external embedding file
raise "use_external_speaker_embedding_file is True, so you need pass a external speaker embedding file, run GE2E-Speaker_Encoder-ExtractSpeakerEmbeddings-by-sample.ipynb or AngularPrototypical-Speaker_Encoder-ExtractSpeakerEmbeddings-by-sample.ipynb notebook in notebooks/ folder"
else: # if start new train and don't use External Embedding file
speaker_mapping = {name: i for i, name in enumerate(speakers)}
speaker_embedding_dim = None
save_speaker_mapping(OUT_PATH, speaker_mapping)
num_speakers = len(speaker_mapping)
print("Training with {} speakers: {}".format(num_speakers,
", ".join(speakers)))
else:
num_speakers = 0
speaker_embedding_dim = None
speaker_mapping = None
model = setup_model(num_chars, num_speakers, c, speaker_embedding_dim)
params = set_weight_decay(model, c.wd)
optimizer = RAdam(params, lr=c.lr, weight_decay=0)
if c.stopnet and c.separate_stopnet:
optimizer_st = RAdam(model.decoder.stopnet.parameters(),
lr=c.lr,
weight_decay=0)
else:
optimizer_st = None
if c.apex_amp_level == "O1":
# pylint: disable=import-outside-toplevel
from apex import amp
model.cuda()
model, optimizer = amp.initialize(model,
optimizer,
opt_level=c.apex_amp_level)
else:
amp = None
# setup criterion
criterion = TacotronLoss(c, stopnet_pos_weight=10.0, ga_sigma=0.4)
if args.restore_path:
checkpoint = torch.load(args.restore_path, map_location='cpu')
try:
# TODO: fix optimizer init, model.cuda() needs to be called before
# optimizer restore
# optimizer.load_state_dict(checkpoint['optimizer'])
if c.reinit_layers:
raise RuntimeError
model.load_state_dict(checkpoint['model'])
except KeyError:
print(" > Partial model initialization.")
model_dict = model.state_dict()
model_dict = set_init_dict(model_dict, checkpoint['model'], c)
# torch.save(model_dict, os.path.join(OUT_PATH, 'state_dict.pt'))
# print("State Dict saved for debug in: ", os.path.join(OUT_PATH, 'state_dict.pt'))
model.load_state_dict(model_dict)
del model_dict
if amp and 'amp' in checkpoint:
amp.load_state_dict(checkpoint['amp'])
for group in optimizer.param_groups:
group['lr'] = c.lr
print(" > Model restored from step %d" % checkpoint['step'],
flush=True)
args.restore_step = checkpoint['step']
else:
args.restore_step = 0
if use_cuda:
model.cuda()
criterion.cuda()
# DISTRUBUTED
if num_gpus > 1:
model = apply_gradient_allreduce(model)
if c.noam_schedule:
scheduler = NoamLR(optimizer,
warmup_steps=c.warmup_steps,
last_epoch=args.restore_step - 1)
else:
scheduler = None
num_params = count_parameters(model)
print("\n > Model has {} parameters".format(num_params), flush=True)
if 'best_loss' not in locals():
best_loss = float('inf')
global_step = args.restore_step
for epoch in range(0, c.epochs):
c_logger.print_epoch_start(epoch, c.epochs)
# set gradual training
if c.gradual_training is not None:
r, c.batch_size = gradual_training_scheduler(global_step, c)
c.r = r
model.decoder.set_r(r)
if c.bidirectional_decoder:
model.decoder_backward.set_r(r)
print("\n > Number of output frames:", model.decoder.r)
train_avg_loss_dict, global_step = train(model, criterion, optimizer,
optimizer_st, scheduler, ap,
global_step, epoch, amp,
speaker_mapping)
eval_avg_loss_dict = evaluate(model, criterion, ap, global_step, epoch,
speaker_mapping)
c_logger.print_epoch_end(epoch, eval_avg_loss_dict)
target_loss = train_avg_loss_dict['avg_postnet_loss']
if c.run_eval:
target_loss = eval_avg_loss_dict['avg_postnet_loss']
best_loss = save_best_model(
target_loss,
best_loss,
model,
optimizer,
global_step,
epoch,
c.r,
OUT_PATH,
amp_state_dict=amp.state_dict() if amp else None)
def get_criterion(self) -> nn.Module:
return TacotronLoss(self.config)
def main(args): # pylint: disable=redefined-outer-name
# pylint: disable=global-variable-undefined
global meta_data_train, meta_data_eval, symbols, phonemes
# Audio processor
ap = AudioProcessor(**c.audio)
if 'characters' in c.keys():
symbols, phonemes = make_symbols(**c.characters)
# DISTRUBUTED
if num_gpus > 1:
init_distributed(args.rank, num_gpus, args.group_id,
c.distributed["backend"], c.distributed["url"])
num_chars = len(phonemes) if c.use_phonemes else len(symbols)
# load data instances
meta_data_train, meta_data_eval = load_meta_data(c.datasets)
# set the portion of the data used for training
if 'train_portion' in c.keys():
meta_data_train = meta_data_train[:int(
len(meta_data_train) * c.train_portion)]
if 'eval_portion' in c.keys():
meta_data_eval = meta_data_eval[:int(
len(meta_data_eval) * c.eval_portion)]
# parse speakers
num_speakers, speaker_embedding_dim, speaker_mapping = parse_speakers(
c, args, meta_data_train, OUT_PATH)
model = setup_model(num_chars, num_speakers, c, speaker_embedding_dim)
# scalers for mixed precision training
scaler = torch.cuda.amp.GradScaler() if c.mixed_precision else None
scaler_st = torch.cuda.amp.GradScaler(
) if c.mixed_precision and c.separate_stopnet else None
params = set_weight_decay(model, c.wd)
optimizer = RAdam(params, lr=c.lr, weight_decay=0)
if c.stopnet and c.separate_stopnet:
optimizer_st = RAdam(model.decoder.stopnet.parameters(),
lr=c.lr,
weight_decay=0)
else:
optimizer_st = None
# setup criterion
criterion = TacotronLoss(c, stopnet_pos_weight=10.0, ga_sigma=0.4)
if args.restore_path:
checkpoint = torch.load(args.restore_path, map_location='cpu')
try:
print(" > Restoring Model.")
model.load_state_dict(checkpoint['model'])
# optimizer restore
print(" > Restoring Optimizer.")
optimizer.load_state_dict(checkpoint['optimizer'])
if "scaler" in checkpoint and c.mixed_precision:
print(" > Restoring AMP Scaler...")
scaler.load_state_dict(checkpoint["scaler"])
if c.reinit_layers:
raise RuntimeError
except KeyError:
print(" > Partial model initialization.")
model_dict = model.state_dict()
model_dict = set_init_dict(model_dict, checkpoint['model'], c)
# torch.save(model_dict, os.path.join(OUT_PATH, 'state_dict.pt'))
# print("State Dict saved for debug in: ", os.path.join(OUT_PATH, 'state_dict.pt'))
model.load_state_dict(model_dict)
del model_dict
for group in optimizer.param_groups:
group['lr'] = c.lr
print(" > Model restored from step %d" % checkpoint['step'],
flush=True)
args.restore_step = checkpoint['step']
else:
args.restore_step = 0
if use_cuda:
model.cuda()
criterion.cuda()
# DISTRUBUTED
if num_gpus > 1:
model = apply_gradient_allreduce(model)
if c.noam_schedule:
scheduler = NoamLR(optimizer,
warmup_steps=c.warmup_steps,
last_epoch=args.restore_step - 1)
else:
scheduler = None
num_params = count_parameters(model)
print("\n > Model has {} parameters".format(num_params), flush=True)
if 'best_loss' not in locals():
best_loss = float('inf')
global_step = args.restore_step
for epoch in range(0, c.epochs):
c_logger.print_epoch_start(epoch, c.epochs)
# set gradual training
if c.gradual_training is not None:
r, c.batch_size = gradual_training_scheduler(global_step, c)
c.r = r
model.decoder.set_r(r)
if c.bidirectional_decoder:
model.decoder_backward.set_r(r)
print("\n > Number of output frames:", model.decoder.r)
train_avg_loss_dict, global_step = train(model, criterion, optimizer,
optimizer_st, scheduler, ap,
global_step, epoch, scaler,
scaler_st, speaker_mapping)
eval_avg_loss_dict = evaluate(model, criterion, ap, global_step, epoch,
speaker_mapping)
c_logger.print_epoch_end(epoch, eval_avg_loss_dict)
target_loss = train_avg_loss_dict['avg_postnet_loss']
if c.run_eval:
target_loss = eval_avg_loss_dict['avg_postnet_loss']
best_loss = save_best_model(
target_loss,
best_loss,
model,
optimizer,
global_step,
epoch,
c.r,
OUT_PATH,
scaler=scaler.state_dict() if c.mixed_precision else None)
def main(args): # pylint: disable=redefined-outer-name
# pylint: disable=global-variable-undefined
global meta_data_train, meta_data_eval, symbols, phonemes
# Audio processor
ap = AudioProcessor(**c.audio)
if 'characters' in c.keys():
symbols, phonemes = make_symbols(**c.characters)
# DISTRUBUTED
if num_gpus > 1:
init_distributed(args.rank, num_gpus, args.group_id,
c.distributed["backend"], c.distributed["url"])
num_chars = len(phonemes) if c.use_phonemes else len(symbols)
# load data instances
meta_data_train, meta_data_eval = load_meta_data(c.datasets)
# set the portion of the data used for training
if 'train_portion' in c.keys():
meta_data_train = meta_data_train[:int(len(meta_data_train) * c.train_portion)]
if 'eval_portion' in c.keys():
meta_data_eval = meta_data_eval[:int(len(meta_data_eval) * c.eval_portion)]
# parse speakers
if c.use_speaker_embedding:
speakers = get_speakers(meta_data_train)
if args.restore_path:
prev_out_path = os.path.dirname(args.restore_path)
speaker_mapping = load_speaker_mapping(prev_out_path)
assert all([speaker in speaker_mapping
for speaker in speakers]), "As of now you, you cannot " \
"introduce new speakers to " \
"a previously trained model."
else:
speaker_mapping = {name: i for i, name in enumerate(speakers)}
save_speaker_mapping(OUT_PATH, speaker_mapping)
num_speakers = len(speaker_mapping)
print("Training with {} speakers: {}".format(num_speakers,
", ".join(speakers)))
else:
num_speakers = 0
model = setup_model(num_chars, num_speakers, c)
params = set_weight_decay(model, c.wd)
optimizer = RAdam(params, lr=c.lr, weight_decay=0)
if c.stopnet and c.separate_stopnet:
optimizer_st = RAdam(model.decoder.stopnet.parameters(),
lr=c.lr,
weight_decay=0)
else:
optimizer_st = None
# setup criterion
criterion = TacotronLoss(c, stopnet_pos_weight=10.0, ga_sigma=0.4)
if args.restore_path:
checkpoint = torch.load(args.restore_path, map_location='cpu')
try:
# TODO: fix optimizer init, model.cuda() needs to be called before
# optimizer restore
# optimizer.load_state_dict(checkpoint['optimizer'])
if c.reinit_layers:
raise RuntimeError
model.load_state_dict(checkpoint['model'])
except:
print(" > Partial model initialization.")
model_dict = model.state_dict()
model_dict = set_init_dict(model_dict, checkpoint['model'], c)
model.load_state_dict(model_dict)
del model_dict
for group in optimizer.param_groups:
group['lr'] = c.lr
print(" > Model restored from step %d" % checkpoint['step'],
flush=True)
args.restore_step = checkpoint['step']
else:
args.restore_step = 0
if use_cuda:
model.cuda()
criterion.cuda()
# DISTRUBUTED
if num_gpus > 1:
model = apply_gradient_allreduce(model)
if c.noam_schedule:
scheduler = NoamLR(optimizer,
warmup_steps=c.warmup_steps,
last_epoch=args.restore_step - 1)
else:
scheduler = None
num_params = count_parameters(model)
print("\n > Model has {} parameters".format(num_params), flush=True)
if 'best_loss' not in locals():
best_loss = float('inf')
global_step = args.restore_step
for epoch in range(0, c.epochs):
c_logger.print_epoch_start(epoch, c.epochs)
# set gradual training
if c.gradual_training is not None:
r, c.batch_size = gradual_training_scheduler(global_step, c)
c.r = r
model.decoder.set_r(r)
if c.bidirectional_decoder:
model.decoder_backward.set_r(r)
print("\n > Number of output frames:", model.decoder.r)
train_avg_loss_dict, global_step = train(model, criterion, optimizer,
optimizer_st, scheduler, ap,
global_step, epoch)
eval_avg_loss_dict = evaluate(model, criterion, ap, global_step, epoch)
c_logger.print_epoch_end(epoch, eval_avg_loss_dict)
target_loss = train_avg_loss_dict['avg_postnet_loss']
if c.run_eval:
target_loss = eval_avg_loss_dict['avg_postnet_loss']
best_loss = save_best_model(target_loss, best_loss, model, optimizer, global_step, epoch, c.r,
OUT_PATH)