def synthesize(speaker, input_file, output_file, params):
print("[Encoding]")
from io_modules.dataset import Dataset
from io_modules.dataset import Encodings
from models.encoder import Encoder
from trainers.encoder import Trainer
encodings = Encodings()
encodings.load('data/models/encoder.encodings')
encoder = Encoder(params, encodings, runtime=True)
encoder.load('data/models/rnn_encoder')
seq = create_lab_input(input_file, speaker)
mgc, att = encoder.generate(seq)
_render_spectrogram(mgc, output_file + '.png')
print("[Vocoding]")
from models.vocoder import Vocoder
from trainers.vocoder import Trainer
vocoder = Vocoder(params, runtime=True)
vocoder.load('data/models/rnn_vocoder')
import time
start = time.time()
signal = vocoder.synthesize(mgc,
batch_size=1000,
temperature=params.temperature)
stop = time.time()
sys.stdout.write(" execution time=" + str(stop - start))
sys.stdout.write('\n')
sys.stdout.flush()
from io_modules.dataset import DatasetIO
dio = DatasetIO()
enc = dio.b16_dec(signal, discreete=True)
dio.write_wave(output_file, enc, params.target_sample_rate)
def phase_4_train_pvocoder(params):
from io_modules.dataset import Dataset
from models.vocoder import Vocoder
from models.vocoder import ParallelVocoder
from trainers.vocoder import Trainer
vocoder_wavenet = Vocoder(params)
sys.stdout.write('Loading wavenet vocoder\n')
vocoder_wavenet.load('data/models/nn_vocoder')
vocoder = ParallelVocoder(params, vocoder_wavenet)
if params.resume:
sys.stdout.write('Resuming from previous checkpoint\n')
vocoder.load('data/models/pnn_vocoder')
trainset = Dataset("data/processed/train")
devset = Dataset("data/processed/dev")
sys.stdout.write('Found ' + str(len(trainset.files)) +
' training files and ' + str(len(devset.files)) +
' development files\n')
trainer = Trainer(vocoder,
trainset,
devset,
target_output_path='data/models/pnn_vocoder')
trainer.start_training(20,
params.batch_size,
params.target_sample_rate,
params=params)
def load_vocoder(params, base_path='data/models'):
from models.vocoder import ParallelVocoder
from models.vocoder import Vocoder
vocoder = Vocoder(params)
vocoder.load('%s/nn_vocoder' % base_path)
pvocoder = ParallelVocoder(params, vocoder=vocoder)
pvocoder.load('%s/pnn_vocoder' % base_path)
return pvocoder
def phase_5_test_vocoder(params):
from io_modules.dataset import Dataset
from models.vocoder import Vocoder
from trainers.vocoder import Trainer
vocoder = Vocoder(params, runtime=True)
vocoder.load('data/models/rnn')
trainset = Dataset("data/processed/train")
devset = Dataset("data/processed/dev")
sys.stdout.write('Found ' + str(len(trainset.files)) + ' training files and ' + str(
len(devset.files)) + ' development files\n')
trainer = Trainer(vocoder, trainset, devset)
trainer.synth_devset(params.batch_size, target_sample_rate=params.target_sample_rate, sample=True,
temperature=0.8)
def phase_2_train_vocoder(params):
from io_modules.dataset import Dataset
from models.vocoder import Vocoder
from trainers.vocoder import Trainer
vocoder = Vocoder(params)
if params.resume:
sys.stdout.write('Resuming from previous checkpoint\n')
vocoder.load('data/models/rnn')
trainset = Dataset("data/processed/train")
devset = Dataset("data/processed/dev")
sys.stdout.write('Found ' + str(len(trainset.files)) + ' training files and ' + str(
len(devset.files)) + ' development files\n')
trainer = Trainer(vocoder, trainset, devset)
trainer.start_training(20, params.batch_size, params.target_sample_rate)
def synthesize(speaker, input_file, output_file, params):
from models.vocoder import device
print(device)
print("[Encoding]")
from io_modules.dataset import Dataset
from io_modules.dataset import Encodings
from models.encoder import Encoder
from trainers.encoder import Trainer
encodings = Encodings()
encodings.load('data/models/encoder.encodings')
encoder = Encoder(params, encodings, runtime=True)
encoder.load('data/models/rnn_encoder')
seq = create_lab_input(input_file, speaker)
mgc, att = encoder.generate(seq)
_render_spectrogram(mgc, output_file + '.png')
print("[Vocoding]")
from models.vocoder import ParallelVocoder
from models.vocoder import Vocoder
vocoder = Vocoder(params)
vocoder.load('data/models/nn_vocoder')
pvocoder = ParallelVocoder(params, vocoder=vocoder)
pvocoder.load('data/models/pnn_vocoder')
import time
start = time.time()
import torch
with torch.no_grad():
signal = pvocoder.synthesize(mgc, batch_size=params.batch_size)
stop = time.time()
sys.stdout.write(" execution time=" + str(stop - start))
sys.stdout.write('\n')
sys.stdout.flush()
from io_modules.dataset import DatasetIO
dio = DatasetIO()
dio.write_wave(output_file, signal / 32768.0, params.target_sample_rate, dtype=signal.dtype)
def phase_7_train_sparse(params):
sys.stdout.write("Starting sparsification for VanillaLSTM\n")
from io_modules.dataset import Dataset
from models.vocoder import Vocoder
from trainers.vocoder import Trainer
vocoder = Vocoder(params, use_sparse_lstm=True)
sys.stdout.write('Resuming from previous checkpoint\n')
vocoder.load('data/models/rnn_vocoder_sparse')
sys.stdout.write("Reading datasets\n")
sys.stdout.flush()
trainset = Dataset("data/processed/train")
devset = Dataset("data/processed/dev")
sys.stdout.write('Found ' + str(len(trainset.files)) +
' training files and ' + str(len(devset.files)) +
' development files\n')
sys.stdout.flush()
trainer = Trainer(vocoder, trainset, devset)
trainer.start_training(20,
params.batch_size,
params.target_sample_rate,
params=params)