Mirrors/Retrieval-based-Voice-Conversion-WebUI
1
0
Fork 0
mirror of https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI.git synced 2025-12-23 23:20:10 +01:00
Code Issues Packages Projects Releases Wiki Activity

Reformat and rewrite _get_name_params (#57)

Browse Source 
* Reformat

* rewrite _get_name_params

* Add workflow for automatic formatting

* Revert "Add workflow for automatic formatting"

This reverts commit 9111c5dbc1.

* revert Retrieval_based_Voice_Conversion_WebUI.ipynb

---------

Co-authored-by: 源文雨 <41315874+fumiama@users.noreply.github.com>
This commit is contained in:
Ftps
2023-04-15 20:44:24 +09:00
committed by GitHub
parent aaa893c4b1
commit c8261b2ccc
45 changed files with 4878 additions and 2456 deletions
Download Patch File Download Diff File Expand all files Collapse all files

567
train_nsf_sim_cache_sid_load_pretrain.py
View File

@@ -1,12 +1,15 @@
import sys,os
now_dir=os.getcwd()
sys.path.append(os.path.join(now_dir,"train"))
import sys, os
now_dir = os.getcwd()
sys.path.append(os.path.join(now_dir, "train"))
import utils
hps = utils.get_hparams()
os.environ["CUDA_VISIBLE_DEVICES"]=hps.gpus.replace("-",",")
n_gpus=len(hps.gpus.split("-"))
os.environ["CUDA_VISIBLE_DEVICES"] = hps.gpus.replace("-", ",")
n_gpus = len(hps.gpus.split("-"))
from random import shuffle
import traceback,json,argparse,itertools,math,torch,pdb
import traceback, json, argparse, itertools, math, torch, pdb
torch.backends.cudnn.deterministic = False
torch.backends.cudnn.benchmark = False
from torch import nn, optim
@@ -20,9 +23,16 @@ from torch.cuda.amp import autocast, GradScaler
from infer_pack import commons
from time import time as ttime
from data_utils import TextAudioLoaderMultiNSFsid,TextAudioLoader, TextAudioCollateMultiNSFsid,TextAudioCollate, DistributedBucketSampler
from data_utils import (
TextAudioLoaderMultiNSFsid,
TextAudioLoader,
TextAudioCollateMultiNSFsid,
TextAudioCollate,
DistributedBucketSampler,
)
from infer_pack.models import (
SynthesizerTrnMs256NSFsid,SynthesizerTrnMs256NSFsid_nono,
SynthesizerTrnMs256NSFsid,
SynthesizerTrnMs256NSFsid_nono,
MultiPeriodDiscriminator,
)
from losses import generator_loss, discriminator_loss, feature_loss, kl_loss
@@ -32,13 +42,11 @@ from mel_processing import mel_spectrogram_torch, spec_to_mel_torch
global_step = 0
def main():
# n_gpus = torch.cuda.device_count()
os.environ["MASTER_ADDR"] = "localhost"
os.environ["MASTER_PORT"] = "5555"
mp.spawn(
run,
nprocs=n_gpus,
@@ -62,13 +70,16 @@ def run(rank, n_gpus, hps):
backend="gloo", init_method="env://", world_size=n_gpus, rank=rank
)
torch.manual_seed(hps.train.seed)
if torch.cuda.is_available(): torch.cuda.set_device(rank)
if torch.cuda.is_available():
torch.cuda.set_device(rank)
if (hps.if_f0 == 1):train_dataset = TextAudioLoaderMultiNSFsid(hps.data.training_files, hps.data)
else:train_dataset = TextAudioLoader(hps.data.training_files, hps.data)
if hps.if_f0 == 1:
train_dataset = TextAudioLoaderMultiNSFsid(hps.data.training_files, hps.data)
else:
train_dataset = TextAudioLoader(hps.data.training_files, hps.data)
train_sampler = DistributedBucketSampler(
train_dataset,
hps.train.batch_size*n_gpus,
hps.train.batch_size * n_gpus,
# [100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1200,1400], # 16s
[100, 200, 300, 400, 500, 600, 700, 800, 900], # 16s
num_replicas=n_gpus,
@@ -77,8 +88,10 @@ def run(rank, n_gpus, hps):
)
# It is possible that dataloader's workers are out of shared memory. Please try to raise your shared memory limit.
# num_workers=8 -> num_workers=4
if (hps.if_f0 == 1):collate_fn = TextAudioCollateMultiNSFsid()
else:collate_fn = TextAudioCollate()
if hps.if_f0 == 1:
collate_fn = TextAudioCollateMultiNSFsid()
else:
collate_fn = TextAudioCollate()
train_loader = DataLoader(
train_dataset,
num_workers=4,
@@ -89,13 +102,26 @@ def run(rank, n_gpus, hps):
persistent_workers=True,
prefetch_factor=8,
)
if(hps.if_f0==1):
net_g = SynthesizerTrnMs256NSFsid(hps.data.filter_length // 2 + 1,hps.train.segment_size // hps.data.hop_length,**hps.model,is_half=hps.train.fp16_run,sr=hps.sample_rate)
if hps.if_f0 == 1:
net_g = SynthesizerTrnMs256NSFsid(
hps.data.filter_length // 2 + 1,
hps.train.segment_size // hps.data.hop_length,
**hps.model,
is_half=hps.train.fp16_run,
sr=hps.sample_rate,
)
else:
net_g = SynthesizerTrnMs256NSFsid_nono(hps.data.filter_length // 2 + 1,hps.train.segment_size // hps.data.hop_length,**hps.model,is_half=hps.train.fp16_run)
if torch.cuda.is_available(): net_g = net_g.cuda(rank)
net_g = SynthesizerTrnMs256NSFsid_nono(
hps.data.filter_length // 2 + 1,
hps.train.segment_size // hps.data.hop_length,
**hps.model,
is_half=hps.train.fp16_run,
)
if torch.cuda.is_available():
net_g = net_g.cuda(rank)
net_d = MultiPeriodDiscriminator(hps.model.use_spectral_norm)
if torch.cuda.is_available(): net_d = net_d.cuda(rank)
if torch.cuda.is_available():
net_d = net_d.cuda(rank)
optim_g = torch.optim.AdamW(
net_g.parameters(),
hps.train.learning_rate,
@@ -110,30 +136,42 @@ def run(rank, n_gpus, hps):
)
# net_g = DDP(net_g, device_ids=[rank], find_unused_parameters=True)
# net_d = DDP(net_d, device_ids=[rank], find_unused_parameters=True)
if torch.cuda.is_available():
if torch.cuda.is_available():
net_g = DDP(net_g, device_ids=[rank])
net_d = DDP(net_d, device_ids=[rank])
else:
net_g = DDP(net_g)
net_d = DDP(net_d)
try:#如果能加载自动resume
_, _, _, epoch_str = utils.load_checkpoint(utils.latest_checkpoint_path(hps.model_dir, "D_*.pth"), net_d, optim_d) # D多半加载没事
try: # 如果能加载自动resume
_, _, _, epoch_str = utils.load_checkpoint(
utils.latest_checkpoint_path(hps.model_dir, "D_*.pth"), net_d, optim_d
) # D多半加载没事
if rank == 0:
logger.info("loaded D")
# _, _, _, epoch_str = utils.load_checkpoint(utils.latest_checkpoint_path(hps.model_dir, "G_*.pth"), net_g, optim_g,load_opt=0)
_, _, _, epoch_str = utils.load_checkpoint(utils.latest_checkpoint_path(hps.model_dir, "G_*.pth"), net_g, optim_g)
_, _, _, epoch_str = utils.load_checkpoint(
utils.latest_checkpoint_path(hps.model_dir, "G_*.pth"), net_g, optim_g
)
global_step = (epoch_str - 1) * len(train_loader)
# epoch_str = 1
# global_step = 0
except:#如果首次不能加载，加载pretrain
except: # 如果首次不能加载，加载pretrain
traceback.print_exc()
epoch_str = 1
global_step = 0
if rank == 0:
logger.info("loaded pretrained %s %s"%(hps.pretrainG,hps.pretrainD))
print(net_g.module.load_state_dict(torch.load(hps.pretrainG,map_location="cpu")["model"]))##测试不加载优化器
print(net_d.module.load_state_dict(torch.load(hps.pretrainD,map_location="cpu")["model"]))
logger.info("loaded pretrained %s %s" % (hps.pretrainG, hps.pretrainD))
print(
net_g.module.load_state_dict(
torch.load(hps.pretrainG, map_location="cpu")["model"]
)
) ##测试不加载优化器
print(
net_d.module.load_state_dict(
torch.load(hps.pretrainD, map_location="cpu")["model"]
)
)
scheduler_g = torch.optim.lr_scheduler.ExponentialLR(
optim_g, gamma=hps.train.lr_decay, last_epoch=epoch_str - 2
@@ -144,7 +182,7 @@ def run(rank, n_gpus, hps):
scaler = GradScaler(enabled=hps.train.fp16_run)
cache=[]
cache = []
for epoch in range(epoch_str, hps.train.epochs + 1):
if rank == 0:
train_and_evaluate(
@@ -157,7 +195,8 @@ def run(rank, n_gpus, hps):
scaler,
[train_loader, None],
logger,
[writer, writer_eval],cache
[writer, writer_eval],
cache,
)
else:
train_and_evaluate(
@@ -170,14 +209,15 @@ def run(rank, n_gpus, hps):
scaler,
[train_loader, None],
None,
None,cache
None,
cache,
)
scheduler_g.step()
scheduler_d.step()
def train_and_evaluate(
rank, epoch, hps, nets, optims, schedulers, scaler, loaders, logger, writers,cache
rank, epoch, hps, nets, optims, schedulers, scaler, loaders, logger, writers, cache
):
net_g, net_d = nets
optim_g, optim_d = optims
@@ -190,168 +230,90 @@ def train_and_evaluate(
net_g.train()
net_d.train()
if(cache==[]or hps.if_cache_data_in_gpu==False):#第一个epoch把cache全部填满训练集
if cache == [] or hps.if_cache_data_in_gpu == False: # 第一个epoch把cache全部填满训练集
# print("caching")
for batch_idx, info in enumerate(train_loader):
if (hps.if_f0 == 1):phone,phone_lengths,pitch,pitchf,spec,spec_lengths,wave,wave_lengths,sid=info
else:phone,phone_lengths,spec,spec_lengths,wave,wave_lengths,sid=info
if hps.if_f0 == 1:
(
phone,
phone_lengths,
pitch,
pitchf,
spec,
spec_lengths,
wave,
wave_lengths,
sid,
) = info
else:
phone, phone_lengths, spec, spec_lengths, wave, wave_lengths, sid = info
if torch.cuda.is_available():
phone, phone_lengths = phone.cuda(rank, non_blocking=True), phone_lengths.cuda(rank, non_blocking=True )
if (hps.if_f0 == 1):pitch,pitchf = pitch.cuda(rank, non_blocking=True),pitchf.cuda(rank, non_blocking=True)
phone, phone_lengths = phone.cuda(
rank, non_blocking=True
), phone_lengths.cuda(rank, non_blocking=True)
if hps.if_f0 == 1:
pitch, pitchf = pitch.cuda(rank, non_blocking=True), pitchf.cuda(
rank, non_blocking=True
)
sid = sid.cuda(rank, non_blocking=True)
spec, spec_lengths = spec.cuda(rank, non_blocking=True), spec_lengths.cuda(rank, non_blocking=True)
wave, wave_lengths = wave.cuda(rank, non_blocking=True), wave_lengths.cuda(rank, non_blocking=True)
if(hps.if_cache_data_in_gpu==True):
if (hps.if_f0 == 1):cache.append((batch_idx, (phone,phone_lengths,pitch,pitchf,spec,spec_lengths,wave,wave_lengths ,sid)))
else:cache.append((batch_idx, (phone,phone_lengths,spec,spec_lengths,wave,wave_lengths ,sid)))
with autocast(enabled=hps.train.fp16_run):
if (hps.if_f0 == 1):y_hat,ids_slice,x_mask,z_mask,(z, z_p, m_p, logs_p, m_q, logs_q) = net_g(phone, phone_lengths, pitch,pitchf, spec, spec_lengths,sid)
else:y_hat,ids_slice,x_mask,z_mask,(z, z_p, m_p, logs_p, m_q, logs_q) = net_g(phone, phone_lengths, spec, spec_lengths,sid)
mel = spec_to_mel_torch(spec,hps.data.filter_length,hps.data.n_mel_channels,hps.data.sampling_rate,hps.data.mel_fmin,hps.data.mel_fmax,)
y_mel = commons.slice_segments(mel, ids_slice, hps.train.segment_size // hps.data.hop_length)
with autocast(enabled=False):
y_hat_mel = mel_spectrogram_torch(
y_hat.float().squeeze(1),
hps.data.filter_length,
hps.data.n_mel_channels,
hps.data.sampling_rate,
hps.data.hop_length,
hps.data.win_length,
hps.data.mel_fmin,
hps.data.mel_fmax,
)
if(hps.train.fp16_run==True):
y_hat_mel=y_hat_mel.half()
wave = commons.slice_segments(
wave, ids_slice * hps.data.hop_length, hps.train.segment_size
) # slice
# Discriminator
y_d_hat_r, y_d_hat_g, _, _ = net_d(wave, y_hat.detach())
with autocast(enabled=False):
loss_disc, losses_disc_r, losses_disc_g = discriminator_loss(
y_d_hat_r, y_d_hat_g
)
optim_d.zero_grad()
scaler.scale(loss_disc).backward()
scaler.unscale_(optim_d)
grad_norm_d = commons.clip_grad_value_(net_d.parameters(), None)
scaler.step(optim_d)
with autocast(enabled=hps.train.fp16_run):
# Generator
y_d_hat_r, y_d_hat_g, fmap_r, fmap_g = net_d(wave, y_hat)
with autocast(enabled=False):
loss_mel = F.l1_loss(y_mel, y_hat_mel) * hps.train.c_mel
loss_kl = kl_loss(z_p, logs_q, m_p, logs_p, z_mask) * hps.train.c_kl
loss_fm = feature_loss(fmap_r, fmap_g)
loss_gen, losses_gen = generator_loss(y_d_hat_g)
loss_gen_all = loss_gen + loss_fm + loss_mel + loss_kl
optim_g.zero_grad()
scaler.scale(loss_gen_all).backward()
scaler.unscale_(optim_g)
grad_norm_g = commons.clip_grad_value_(net_g.parameters(), None)
scaler.step(optim_g)
scaler.update()
if rank == 0:
if global_step % hps.train.log_interval == 0:
lr = optim_g.param_groups[0]["lr"]
logger.info(
"Train Epoch: {} [{:.0f}%]".format(
epoch, 100.0 * batch_idx / len(train_loader)
spec, spec_lengths = spec.cuda(
rank, non_blocking=True
), spec_lengths.cuda(rank, non_blocking=True)
wave, wave_lengths = wave.cuda(
rank, non_blocking=True
), wave_lengths.cuda(rank, non_blocking=True)
if hps.if_cache_data_in_gpu == True:
if hps.if_f0 == 1:
cache.append(
(
batch_idx,
(
phone,
phone_lengths,
pitch,
pitchf,
spec,
spec_lengths,
wave,
wave_lengths,
sid,
),
)
)
# Amor For Tensorboard display
if loss_mel > 50:
loss_mel = 50
if loss_kl > 5:
loss_kl = 5
logger.info([global_step, lr])
logger.info(
f"loss_disc={loss_disc:.3f}, loss_gen={loss_gen:.3f}, loss_fm={loss_fm:.3f},loss_mel={loss_mel:.3f}, loss_kl={loss_kl:.3f}"
else:
cache.append(
(
batch_idx,
(
phone,
phone_lengths,
spec,
spec_lengths,
wave,
wave_lengths,
sid,
),
)
)
scalar_dict = {
"loss/g/total": loss_gen_all,
"loss/d/total": loss_disc,
"learning_rate": lr,
"grad_norm_d": grad_norm_d,
"grad_norm_g": grad_norm_g,
}
scalar_dict.update(
{"loss/g/fm": loss_fm, "loss/g/mel": loss_mel, "loss/g/kl": loss_kl}
)
scalar_dict.update(
{"loss/g/{}".format(i): v for i, v in enumerate(losses_gen)}
)
scalar_dict.update(
{"loss/d_r/{}".format(i): v for i, v in enumerate(losses_disc_r)}
)
scalar_dict.update(
{"loss/d_g/{}".format(i): v for i, v in enumerate(losses_disc_g)}
)
image_dict = {
"slice/mel_org": utils.plot_spectrogram_to_numpy(
y_mel[0].data.cpu().numpy()
),
"slice/mel_gen": utils.plot_spectrogram_to_numpy(
y_hat_mel[0].data.cpu().numpy()
),
"all/mel": utils.plot_spectrogram_to_numpy(
mel[0].data.cpu().numpy()
),
}
utils.summarize(
writer=writer,
global_step=global_step,
images=image_dict,
scalars=scalar_dict,
)
global_step += 1
# if global_step % hps.train.eval_interval == 0:
if epoch % hps.save_every_epoch == 0 and rank == 0:
if(hps.if_latest==0):
utils.save_checkpoint(
net_g,
optim_g,
hps.train.learning_rate,
epoch,
os.path.join(hps.model_dir, "G_{}.pth".format(global_step)),
)
utils.save_checkpoint(
net_d,
optim_d,
hps.train.learning_rate,
epoch,
os.path.join(hps.model_dir, "D_{}.pth".format(global_step)),
)
else:
utils.save_checkpoint(
net_g,
optim_g,
hps.train.learning_rate,
epoch,
os.path.join(hps.model_dir, "G_{}.pth".format(2333333)),
)
utils.save_checkpoint(
net_d,
optim_d,
hps.train.learning_rate,
epoch,
os.path.join(hps.model_dir, "D_{}.pth".format(2333333)),
)
else:#后续的epoch直接使用打乱的cache
shuffle(cache)
# print("using cache")
for batch_idx, info in cache:
if (hps.if_f0 == 1):phone,phone_lengths,pitch,pitchf,spec,spec_lengths,wave,wave_lengths,sid=info
else:phone,phone_lengths,spec,spec_lengths,wave,wave_lengths,sid=info
with autocast(enabled=hps.train.fp16_run):
if (hps.if_f0 == 1):y_hat,ids_slice,x_mask,z_mask,(z, z_p, m_p, logs_p, m_q, logs_q) = net_g(phone, phone_lengths, pitch,pitchf, spec, spec_lengths,sid)
else:y_hat,ids_slice,x_mask,z_mask,(z, z_p, m_p, logs_p, m_q, logs_q) = net_g(phone, phone_lengths, spec, spec_lengths,sid)
if hps.if_f0 == 1:
(
y_hat,
ids_slice,
x_mask,
z_mask,
(z, z_p, m_p, logs_p, m_q, logs_q),
) = net_g(
phone, phone_lengths, pitch, pitchf, spec, spec_lengths, sid
)
else:
(
y_hat,
ids_slice,
x_mask,
z_mask,
(z, z_p, m_p, logs_p, m_q, logs_q),
) = net_g(phone, phone_lengths, spec, spec_lengths, sid)
mel = spec_to_mel_torch(
spec,
hps.data.filter_length,
@@ -374,8 +336,200 @@ def train_and_evaluate(
hps.data.mel_fmin,
hps.data.mel_fmax,
)
if(hps.train.fp16_run==True):
y_hat_mel=y_hat_mel.half()
if hps.train.fp16_run == True:
y_hat_mel = y_hat_mel.half()
wave = commons.slice_segments(
wave, ids_slice * hps.data.hop_length, hps.train.segment_size
) # slice
# Discriminator
y_d_hat_r, y_d_hat_g, _, _ = net_d(wave, y_hat.detach())
with autocast(enabled=False):
loss_disc, losses_disc_r, losses_disc_g = discriminator_loss(
y_d_hat_r, y_d_hat_g
)
optim_d.zero_grad()
scaler.scale(loss_disc).backward()
scaler.unscale_(optim_d)
grad_norm_d = commons.clip_grad_value_(net_d.parameters(), None)
scaler.step(optim_d)
with autocast(enabled=hps.train.fp16_run):
# Generator
y_d_hat_r, y_d_hat_g, fmap_r, fmap_g = net_d(wave, y_hat)
with autocast(enabled=False):
loss_mel = F.l1_loss(y_mel, y_hat_mel) * hps.train.c_mel
loss_kl = kl_loss(z_p, logs_q, m_p, logs_p, z_mask) * hps.train.c_kl
loss_fm = feature_loss(fmap_r, fmap_g)
loss_gen, losses_gen = generator_loss(y_d_hat_g)
loss_gen_all = loss_gen + loss_fm + loss_mel + loss_kl
optim_g.zero_grad()
scaler.scale(loss_gen_all).backward()
scaler.unscale_(optim_g)
grad_norm_g = commons.clip_grad_value_(net_g.parameters(), None)
scaler.step(optim_g)
scaler.update()
if rank == 0:
if global_step % hps.train.log_interval == 0:
lr = optim_g.param_groups[0]["lr"]
logger.info(
"Train Epoch: {} [{:.0f}%]".format(
epoch, 100.0 * batch_idx / len(train_loader)
)
)
# Amor For Tensorboard display
if loss_mel > 50:
loss_mel = 50
if loss_kl > 5:
loss_kl = 5
logger.info([global_step, lr])
logger.info(
f"loss_disc={loss_disc:.3f}, loss_gen={loss_gen:.3f}, loss_fm={loss_fm:.3f},loss_mel={loss_mel:.3f}, loss_kl={loss_kl:.3f}"
)
scalar_dict = {
"loss/g/total": loss_gen_all,
"loss/d/total": loss_disc,
"learning_rate": lr,
"grad_norm_d": grad_norm_d,
"grad_norm_g": grad_norm_g,
}
scalar_dict.update(
{
"loss/g/fm": loss_fm,
"loss/g/mel": loss_mel,
"loss/g/kl": loss_kl,
}
)
scalar_dict.update(
{"loss/g/{}".format(i): v for i, v in enumerate(losses_gen)}
)
scalar_dict.update(
{
"loss/d_r/{}".format(i): v
for i, v in enumerate(losses_disc_r)
}
)
scalar_dict.update(
{
"loss/d_g/{}".format(i): v
for i, v in enumerate(losses_disc_g)
}
)
image_dict = {
"slice/mel_org": utils.plot_spectrogram_to_numpy(
y_mel[0].data.cpu().numpy()
),
"slice/mel_gen": utils.plot_spectrogram_to_numpy(
y_hat_mel[0].data.cpu().numpy()
),
"all/mel": utils.plot_spectrogram_to_numpy(
mel[0].data.cpu().numpy()
),
}
utils.summarize(
writer=writer,
global_step=global_step,
images=image_dict,
scalars=scalar_dict,
)
global_step += 1
# if global_step % hps.train.eval_interval == 0:
if epoch % hps.save_every_epoch == 0 and rank == 0:
if hps.if_latest == 0:
utils.save_checkpoint(
net_g,
optim_g,
hps.train.learning_rate,
epoch,
os.path.join(hps.model_dir, "G_{}.pth".format(global_step)),
)
utils.save_checkpoint(
net_d,
optim_d,
hps.train.learning_rate,
epoch,
os.path.join(hps.model_dir, "D_{}.pth".format(global_step)),
)
else:
utils.save_checkpoint(
net_g,
optim_g,
hps.train.learning_rate,
epoch,
os.path.join(hps.model_dir, "G_{}.pth".format(2333333)),
)
utils.save_checkpoint(
net_d,
optim_d,
hps.train.learning_rate,
epoch,
os.path.join(hps.model_dir, "D_{}.pth".format(2333333)),
)
else: # 后续的epoch直接使用打乱的cache
shuffle(cache)
# print("using cache")
for batch_idx, info in cache:
if hps.if_f0 == 1:
(
phone,
phone_lengths,
pitch,
pitchf,
spec,
spec_lengths,
wave,
wave_lengths,
sid,
) = info
else:
phone, phone_lengths, spec, spec_lengths, wave, wave_lengths, sid = info
with autocast(enabled=hps.train.fp16_run):
if hps.if_f0 == 1:
(
y_hat,
ids_slice,
x_mask,
z_mask,
(z, z_p, m_p, logs_p, m_q, logs_q),
) = net_g(
phone, phone_lengths, pitch, pitchf, spec, spec_lengths, sid
)
else:
(
y_hat,
ids_slice,
x_mask,
z_mask,
(z, z_p, m_p, logs_p, m_q, logs_q),
) = net_g(phone, phone_lengths, spec, spec_lengths, sid)
mel = spec_to_mel_torch(
spec,
hps.data.filter_length,
hps.data.n_mel_channels,
hps.data.sampling_rate,
hps.data.mel_fmin,
hps.data.mel_fmax,
)
y_mel = commons.slice_segments(
mel, ids_slice, hps.train.segment_size // hps.data.hop_length
)
with autocast(enabled=False):
y_hat_mel = mel_spectrogram_torch(
y_hat.float().squeeze(1),
hps.data.filter_length,
hps.data.n_mel_channels,
hps.data.sampling_rate,
hps.data.hop_length,
hps.data.win_length,
hps.data.mel_fmin,
hps.data.mel_fmax,
)
if hps.train.fp16_run == True:
y_hat_mel = y_hat_mel.half()
wave = commons.slice_segments(
wave, ids_slice * hps.data.hop_length, hps.train.segment_size
) # slice
@@ -435,17 +589,27 @@ def train_and_evaluate(
"grad_norm_g": grad_norm_g,
}
scalar_dict.update(
{"loss/g/fm": loss_fm, "loss/g/mel": loss_mel, "loss/g/kl": loss_kl}
{
"loss/g/fm": loss_fm,
"loss/g/mel": loss_mel,
"loss/g/kl": loss_kl,
}
)
scalar_dict.update(
{"loss/g/{}".format(i): v for i, v in enumerate(losses_gen)}
)
scalar_dict.update(
{"loss/d_r/{}".format(i): v for i, v in enumerate(losses_disc_r)}
{
"loss/d_r/{}".format(i): v
for i, v in enumerate(losses_disc_r)
}
)
scalar_dict.update(
{"loss/d_g/{}".format(i): v for i, v in enumerate(losses_disc_g)}
{
"loss/d_g/{}".format(i): v
for i, v in enumerate(losses_disc_g)
}
)
image_dict = {
"slice/mel_org": utils.plot_spectrogram_to_numpy(
@@ -467,7 +631,7 @@ def train_and_evaluate(
global_step += 1
# if global_step % hps.train.eval_interval == 0:
if epoch % hps.save_every_epoch == 0 and rank == 0:
if(hps.if_latest==0):
if hps.if_latest == 0:
utils.save_checkpoint(
net_g,
optim_g,
@@ -498,15 +662,20 @@ def train_and_evaluate(
os.path.join(hps.model_dir, "D_{}.pth".format(2333333)),
)
if rank == 0:
logger.info("====> Epoch: {}".format(epoch))
if(epoch>=hps.total_epoch and rank == 0):
if epoch >= hps.total_epoch and rank == 0:
logger.info("Training is done. The program is closed.")
from process_ckpt import savee#def savee(ckpt,sr,if_f0,name,epoch):
if hasattr(net_g, 'module'):ckpt = net_g.module.state_dict()
else:ckpt = net_g.state_dict()
logger.info("saving final ckpt:%s"%(savee(ckpt,hps.sample_rate,hps.if_f0,hps.name,epoch)))
from process_ckpt import savee # def savee(ckpt,sr,if_f0,name,epoch):
if hasattr(net_g, "module"):
ckpt = net_g.module.state_dict()
else:
ckpt = net_g.state_dict()
logger.info(
"saving final ckpt:%s"
% (savee(ckpt, hps.sample_rate, hps.if_f0, hps.name, epoch))
)
os._exit(2333333)