def load_checkpoint(checkpoint_path, model, optimizer):
assert os.path.isfile(checkpoint_path)
print("Loading checkpoint '{}'".format(checkpoint_path))
checkpoint_dict = torch.load(checkpoint_path, map_location='cpu')
# 显示模型结构
# from torchkeras import summary
# summary(checkpoint_dict, input_shape=(145,512))
# print("checkpoint_dict的内容:",checkpoint_dict)
# print("checkpoint_dict的类型:",type(checkpoint_dict))
# with open("./checkpoint_dict.txt",'w') as fw:
# fw.write(str(checkpoint_dict))
# 原始的模型导入不成功
# model.load_state_dict(checkpoint_dict['state_dict']) # 原始
# learning_rate = checkpoint_dict['learning_rate'] # 原始
# iteration = checkpoint_dict['iteration'] # 原始
# optimizer.load_state_dict(checkpoint_dict['optimizer']) # 原始
model.load_state_dict(checkpoint_dict.state_dict())
hparams = create_hparams(args.hparams_json, level=args.hparams_level)
learning_rate = hparams.learning_rate
# print("learning_rate的内容:",learning_rate)
iteration = hparams.iters_per_checkpoint
# print("iteration的内容:",iteration)
optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate,
weight_decay=hparams.weight_decay)
# print("optimizer的内容:",optimizer)
print("Loaded checkpoint '{}' from iteration {}".format(
checkpoint_path, iteration))
return model, optimizer, learning_rate, iteration
from mellotron.hparams import create_hparams
from mellotron.train import train, json_dump, parse_args
args = parse_args()
os.environ["CUDA_VISIBLE_DEVICES"] = args.cuda
if __name__ == '__main__':
try:
from setproctitle import setproctitle
setproctitle('zhrtvc-mellotron-train')
except ImportError:
pass
hparams = create_hparams(args.hparams)
torch.backends.cudnn.enabled = hparams.cudnn_enabled
torch.backends.cudnn.benchmark = hparams.cudnn_benchmark
print("FP16 Run:", hparams.fp16_run)
print("Dynamic Loss Scaling:", hparams.dynamic_loss_scaling)
print("Distributed Run:", hparams.distributed_run)
print("cuDNN Enabled:", hparams.cudnn_enabled)
print("cuDNN Benchmark:", hparams.cudnn_benchmark)
meta_folder = os.path.join(args.output_directory, 'metadata')
os.makedirs(meta_folder, exist_ok=True)
path = os.path.join(meta_folder, "args.json")
obj = args.__dict__
import torch
import aukit
import unidecode
import yaml
import librosa
from waveglow import inference as waveglow
from melgan import inference as melgan
from mellotron import inference as mellotron
from utils.argutils import locals2dict
from mellotron.layers import TacotronSTFT
from mellotron.hparams import create_hparams
# 用griffinlim声码器
_hparams = create_hparams()
_stft = TacotronSTFT(_hparams.filter_length, _hparams.hop_length,
_hparams.win_length, _hparams.n_mel_channels,
_hparams.sampling_rate, _hparams.mel_fmin,
_hparams.mel_fmax)
_use_waveglow = 0
_device = 'cuda' if torch.cuda.is_available() else 'cpu'
filename_formatter_re = re.compile(r'[\s\\/:*?"<>|\']+')
def plot_mel_alignment_gate_audio(mel,
alignment,
gate,
audio,
import yaml
import torch
from mellotron.hparams import create_hparams
from mellotron.train import train, json_dump, yaml_dump
if __name__ == '__main__':
try:
from setproctitle import setproctitle
setproctitle('zhrtvc-mellotron-train')
except ImportError:
pass
hparams = create_hparams(args.hparams_json, level=args.hparams_level)
torch.backends.cudnn.enabled = hparams.cudnn_enabled
torch.backends.cudnn.benchmark = hparams.cudnn_benchmark
print("FP16 Run:", hparams.fp16_run)
print("Dynamic Loss Scaling:", hparams.dynamic_loss_scaling)
print("Distributed Run:", hparams.distributed_run)
print("cuDNN Enabled:", hparams.cudnn_enabled)
print("cuDNN Benchmark:", hparams.cudnn_benchmark)
meta_folder = os.path.join(args.output_directory, 'metadata')
os.makedirs(meta_folder, exist_ok=True)
stem_path = os.path.join(meta_folder, "args")
obj = args.__dict__
from mellotron.hparams import create_hparams
from mellotron.train import train, json_dump, parse_args, yaml_dump
args = parse_args()
os.environ["CUDA_VISIBLE_DEVICES"] = args.cuda
if __name__ == '__main__':
try:
from setproctitle import setproctitle
setproctitle('zhrtvc-mellotron-train')
except ImportError:
pass
hparams = create_hparams(args.hparams_json)
torch.backends.cudnn.enabled = hparams.cudnn_enabled
torch.backends.cudnn.benchmark = hparams.cudnn_benchmark
print("FP16 Run:", hparams.fp16_run)
print("Dynamic Loss Scaling:", hparams.dynamic_loss_scaling)
print("Distributed Run:", hparams.distributed_run)
print("cuDNN Enabled:", hparams.cudnn_enabled)
print("cuDNN Benchmark:", hparams.cudnn_benchmark)
meta_folder = os.path.join(args.output_directory, 'metadata')
os.makedirs(meta_folder, exist_ok=True)
stem_path = os.path.join(meta_folder, "args")
obj = args.__dict__