testadmin
/
ParakeetEricRoss
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

add profiling

This commit is contained in:
chenfeiyu 2021-06-16 09:40:47 +00:00
parent bbbe5a8b50
commit 8dbcc9bccb
5 changed files with 19 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
examples/parallelwave_gan/baker/preprocess.py
View File

@ -16,12 +16,9 @@ from typing import List, Dict, Any
import soundfile as sf import soundfile as sf
import librosa import librosa
import numpy as np import numpy as np
from config import get_cfg_default
import argparse import argparse
import yaml import yaml
import json import json
import dacite
import dataclasses
import concurrent.futures import concurrent.futures
from concurrent.futures import ProcessPoolExecutor, ThreadPoolExecutor from concurrent.futures import ProcessPoolExecutor, ThreadPoolExecutor
from pathlib import Path from pathlib import Path
@ -30,6 +27,8 @@ from operator import itemgetter
from praatio import tgio from praatio import tgio
import logging import logging
from config import get_cfg_default
def logmelfilterbank(audio, def logmelfilterbank(audio,
sr, sr,
@ -229,7 +228,7 @@ def main():
if args.verbose > 1: if args.verbose > 1:
print(vars(args)) print(vars(args))
print(yaml.dump(dataclasses.asdict(C))) print(C)
root_dir = Path(args.rootdir) root_dir = Path(args.rootdir)
dumpdir = Path(args.dumpdir) dumpdir = Path(args.dumpdir)

16
examples/parallelwave_gan/baker/pwg_updater.py
View File

@ -13,6 +13,7 @@
# limitations under the License. # limitations under the License.
import paddle import paddle
from paddle.fluid.core import _cuda_synchronize
from timer import timer from timer import timer
from parakeet.datasets.data_table import DataTable from parakeet.datasets.data_table import DataTable
@ -61,12 +62,15 @@ class PWGUpdater(UpdaterBase):
self.train_iterator = iter(self.train_dataloader) self.train_iterator = iter(self.train_dataloader)
def update_core(self): def update_core(self):
place = paddle.fluid.framework._current_expected_place()
with timer() as t: with timer() as t:
_cuda_synchronize(place)
try: try:
batch = next(self.train_iterator) batch = next(self.train_iterator)
except StopIteration: except StopIteration:
self.train_iterator = iter(self.train_dataloader) self.train_iterator = iter(self.train_dataloader)
batch = next(self.train_iterator) batch = next(self.train_iterator)
_cuda_synchronize(place)
print(f"Loading a batch takes {t.elapse}s") print(f"Loading a batch takes {t.elapse}s")
wav, mel = batch wav, mel = batch
@ -75,13 +79,17 @@ class PWGUpdater(UpdaterBase):
noise = paddle.randn(wav.shape) noise = paddle.randn(wav.shape)
with timer() as t: with timer() as t:
_cuda_synchronize(place)
wav_ = self.generator(noise, mel) wav_ = self.generator(noise, mel)
_cuda_synchronize(place)
print(f"Generator takes {t.elapse}s") print(f"Generator takes {t.elapse}s")
## Multi-resolution stft loss ## Multi-resolution stft loss
with timer() as t: with timer() as t:
_cuda_synchronize(place)
sc_loss, mag_loss = self.criterion_stft( sc_loss, mag_loss = self.criterion_stft(
wav_.squeeze(1), wav.squeeze(1)) wav_.squeeze(1), wav.squeeze(1))
_cuda_synchronize(place)
print(f"Multi-resolution STFT loss takes {t.elapse}s") print(f"Multi-resolution STFT loss takes {t.elapse}s")
report("train/spectral_convergence_loss", float(sc_loss)) report("train/spectral_convergence_loss", float(sc_loss))
@ -91,24 +99,30 @@ class PWGUpdater(UpdaterBase):
## Adversarial loss ## Adversarial loss
if self.state.iteration > self.discriminator_train_start_steps: if self.state.iteration > self.discriminator_train_start_steps:
with timer() as t: with timer() as t:
_cuda_synchronize(place)
p_ = self.discriminator(wav_) p_ = self.discriminator(wav_)
adv_loss = self.criterion_mse(p_, paddle.ones_like(p_)) adv_loss = self.criterion_mse(p_, paddle.ones_like(p_))
_cuda_synchronize(place)
print(f"Discriminator and adversarial loss takes {t.elapse}s") print(f"Discriminator and adversarial loss takes {t.elapse}s")
report("train/adversarial_loss", float(adv_loss)) report("train/adversarial_loss", float(adv_loss))
gen_loss += self.lambda_adv * adv_loss gen_loss += self.lambda_adv * adv_loss
report("train/generator_loss", float(gen_loss)) report("train/generator_loss", float(gen_loss))
with timer() as t: with timer() as t:
_cuda_synchronize(place)
self.optimizer_g.clear_grad() self.optimizer_g.clear_grad()
gen_loss.backward() gen_loss.backward()
_cuda_synchronize(place)
print(f"Backward takes {t.elapse}s.") print(f"Backward takes {t.elapse}s.")
with timer() as t: with timer() as t:
_cuda_synchronize(place)
self.optimizer_g.step() self.optimizer_g.step()
self.scheduler_g.step() self.scheduler_g.step()
_cuda_synchronize(place)
print(f"Update takes {t.elapse}s.") print(f"Update takes {t.elapse}s.")
# Disctiminator # Disctiminator
if self.state.iteration > self.discriminator_train_start_steps: if self.state.iteration > self.discriminator_train_start_steps:
with paddle.no_grad(): with paddle.no_grad():
wav_ = self.generator(noise, mel) wav_ = self.generator(noise, mel)

1
examples/parallelwave_gan/baker/train.py
View File

@ -20,7 +20,6 @@ import dataclasses
from pathlib import Path from pathlib import Path
import yaml import yaml
import dacite
import json import json
import paddle import paddle
import numpy as np import numpy as np

2
parakeet/modules/stft_loss.py
View File

@ -64,7 +64,7 @@ class STFTLoss(nn.Layer):
fft_size=1024, fft_size=1024,
shift_size=120, shift_size=120,
win_length=600, win_length=600,
window="hann_window"): window="hann"):
"""Initialize STFT loss module.""" """Initialize STFT loss module."""
super().__init__() super().__init__()
self.fft_size = fft_size self.fft_size = fft_size

1
setup.py
View File

@ -64,7 +64,6 @@ setup_info = dict(
'scipy', 'scipy',
'pandas', 'pandas',
'sox', 'sox',
# 'opencc',
'soundfile', 'soundfile',
'g2p_en', 'g2p_en',
'yacs', 'yacs',