will run on windows (#75)

scripts/check_silence.py

@@ -28,7 +28,6 @@
 parser.add_argument('--speaker_yml', type=Path)
 parser.add_argument('--pad_second', type=float, default=base_acoustic_param.pad_second)
 arguments = parser.parse_args()
-pprint(vars(arguments))
 
 # read parameters from speaker yml
 sconf = SpeakerYML(arguments.speaker_yml)
@@ -78,6 +77,8 @@ def calc_score(path: Path):
 
 
 def main():
+    pprint(vars(arguments))
+
     paths = [Path(p) for p in sorted(glob.glob(arguments.input_wave_glob))]
     pool = multiprocessing.Pool()
     it = pool.imap(calc_score, paths)

scripts/convert_acoustic_feature.py

@@ -23,7 +23,6 @@
 parser.add_argument('--ignore_feature', nargs='+', default=['sp', 'ap'])
 parser.add_argument('--enable_overwrite', action='store_true')
 arguments = parser.parse_args()
-pprint(vars(arguments))
 
 
 def convert_feature(path: Path, acoustic_converter: AcousticConverter):
@@ -40,6 +39,8 @@ def convert_feature(path: Path, acoustic_converter: AcousticConverter):
 
 
 def main():
+    pprint(vars(arguments))
+
     arguments.output.mkdir(exist_ok=True)
     save_arguments(arguments, arguments.output / 'arguments.json')
 

scripts/extract_acoustic_feature.py

@@ -34,7 +34,6 @@
 parser.add_argument('--ignore_feature', nargs='+', default=['sp', 'ap'])
 parser.add_argument('--enable_overwrite', action='store_true')
 arguments = parser.parse_args()
-pprint(vars(arguments))
 
 
 def generate_feature(path: Path):
@@ -85,6 +84,8 @@ def generate_feature(path: Path):
 
 
 def main():
+    pprint(vars(arguments))
+
     arguments.output.mkdir(exist_ok=True)
     save_arguments(arguments, arguments.output / 'arguments.json')
 

scripts/extract_align_indexes.py

@@ -23,7 +23,6 @@
 parser.add_argument('--ignore_feature', nargs='+', default=('feature1', 'feature2'))
 parser.add_argument('--enable_overwrite', action='store_true')
 arguments = parser.parse_args()
-pprint(vars(arguments))
 
 
 def generate_align_indexes(pair_path: Tuple[Path, Path]):
@@ -45,6 +44,8 @@ def generate_align_indexes(pair_path: Tuple[Path, Path]):
 
 
 def main():
+    pprint(vars(arguments))
+
     arguments.output.mkdir(exist_ok=True)
     save_arguments(arguments, arguments.output / 'arguments.json')
 

scripts/extract_converted_spectrogram_pair.py

@@ -23,7 +23,6 @@
 parser.add_argument('--output', '-o', type=Path)
 parser.add_argument('--enable_overwrite', action='store_true')
 arguments = parser.parse_args()
-pprint(vars(arguments))
 
 input_glob: Path = arguments.input_glob
 target_glob: Path = arguments.target_glob
@@ -53,6 +52,8 @@ def generate_file(pair_path: Tuple[Path, Path, Path]):
 
 
 def main():
+    pprint(vars(arguments))
+
     output.mkdir(exist_ok=True)
     save_arguments(arguments, output / 'arguments.json')
 

scripts/extract_f0_statistics.py

@@ -13,7 +13,6 @@
 parser.add_argument('--input_glob', '-i')
 parser.add_argument('--output', '-o', type=Path)
 arguments = parser.parse_args()
-pprint(vars(arguments))
 
 
 def load_f0(path: Path):
@@ -22,6 +21,8 @@ def load_f0(path: Path):
 
 
 def main():
+    pprint(vars(arguments))
+
     paths = [Path(p) for p in sorted(glob.glob(arguments.input_glob))]
 
     pool = multiprocessing.Pool()

scripts/extract_gmm_align_indexes.py

@@ -43,7 +43,6 @@
 parser.add_argument('--ignore_feature', nargs='+', default=('feature1', 'feature2'))
 parser.add_argument('--enable_overwrite', action='store_true')
 arguments = parser.parse_args()
-pprint(vars(arguments))
 
 # read parameters from speaker yml
 sconf1 = SpeakerYML(arguments.org_yml)
@@ -130,6 +129,8 @@ def generate_align_indexes(pair_path: Tuple[Path, Path]):
 
 
 def main():
+    pprint(vars(arguments))
+
     arguments.output.mkdir(exist_ok=True)
     save_arguments(arguments, arguments.output / 'arguments.json')
 

scripts/generate_aligned_wave.py

@@ -31,7 +31,6 @@
 parser.add_argument('--alpha', type=float, default=base_acoustic_param.alpha)
 parser.add_argument('--disable_overwrite', action='store_true')
 arguments = parser.parse_args()
-pprint(vars(arguments))
 
 
 def generate_aligned_wave(
@@ -72,6 +71,8 @@ def generate_aligned_wave(
 
 
 def main():
+    pprint(vars(arguments))
+
     arguments.output.mkdir(exist_ok=True)
     save_arguments(arguments, arguments.output / 'arguments.json')
 

train.py

@@ -20,86 +20,87 @@
 from yukarin.model import create
 from yukarin.updater import Updater
 
-parser = argparse.ArgumentParser()
-parser.add_argument('config_json_path', type=Path)
-parser.add_argument('output', type=Path)
-arguments = parser.parse_args()
-
-config = create_from_json(arguments.config_json_path)
-arguments.output.mkdir(exist_ok=True)
-config.save_as_json((arguments.output / 'config.json').absolute())
-
-# model
-if config.train.gpu >= 0:
-    cuda.get_device_from_id(config.train.gpu).use()
-predictor, discriminator = create(config.model)
-models = {
-    'predictor': predictor,
-    'discriminator': discriminator,
-}
-
-if config.train.pretrained_model is not None:
-    serializers.load_npz(str(config.train.pretrained_model), predictor)
-
-# dataset
-dataset = create_dataset(config.dataset)
-train_iter = MultiprocessIterator(dataset['train'], config.train.batchsize)
-test_iter = MultiprocessIterator(dataset['test'], config.train.batchsize, repeat=False, shuffle=False)
-train_eval_iter = MultiprocessIterator(dataset['train_eval'], config.train.batchsize, repeat=False, shuffle=False)
-
-
-# optimizer
-def create_optimizer(model):
-    cp: Dict[str, Any] = copy(config.train.optimizer)
-    n = cp.pop('name').lower()
-
-    if n == 'adam':
-        optimizer = optimizers.Adam(**cp)
-    elif n == 'sgd':
-        optimizer = optimizers.SGD(**cp)
-    else:
-        raise ValueError(n)
-
-    optimizer.setup(model)
-    return optimizer
-
-
-opts = {key: create_optimizer(model) for key, model in models.items()}
-
-# updater
-converter = partial(convert.concat_examples, padding=0)
-updater = Updater(
-    loss_config=config.loss,
-    predictor=predictor,
-    discriminator=discriminator,
-    device=config.train.gpu,
-    iterator=train_iter,
-    optimizer=opts,
-    converter=converter,
-)
-
-# trainer
-trigger_log = (config.train.log_iteration, 'iteration')
-trigger_snapshot = (config.train.snapshot_iteration, 'iteration')
-trigger_stop = (config.train.stop_iteration, 'iteration') if config.train.stop_iteration is not None else None
-
-trainer = training.Trainer(updater, stop_trigger=trigger_stop, out=arguments.output)
-tb_writer = SummaryWriter(Path(arguments.output))
-
-ext = extensions.Evaluator(test_iter, models, converter, device=config.train.gpu, eval_func=updater.forward)
-trainer.extend(ext, name='test', trigger=trigger_log)
-ext = extensions.Evaluator(train_eval_iter, models, converter, device=config.train.gpu, eval_func=updater.forward)
-trainer.extend(ext, name='train', trigger=trigger_log)
-
-trainer.extend(extensions.dump_graph('predictor/loss'))
-
-ext = extensions.snapshot_object(predictor, filename='predictor_{.updater.iteration}.npz')
-trainer.extend(ext, trigger=trigger_snapshot)
-
-trainer.extend(extensions.LogReport(trigger=trigger_log))
-trainer.extend(TensorBoardReport(writer=tb_writer), trigger=trigger_log)
-
-if trigger_stop is not None:
-    trainer.extend(extensions.ProgressBar(trigger_stop))
-
-trainer.run()
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('config_json_path', type=Path)
+    parser.add_argument('output', type=Path)
+    arguments = parser.parse_args()
+
+    config = create_from_json(arguments.config_json_path)
+    arguments.output.mkdir(exist_ok=True)
+    config.save_as_json((arguments.output / 'config.json').absolute())
+
+    # model
+    if config.train.gpu >= 0:
+        cuda.get_device_from_id(config.train.gpu).use()
+    predictor, discriminator = create(config.model)
+    models = {
+        'predictor': predictor,
+        'discriminator': discriminator,
+    }
+
+    if config.train.pretrained_model is not None:
+        serializers.load_npz(str(config.train.pretrained_model), predictor)
+
+    # dataset
+    dataset = create_dataset(config.dataset)
+    train_iter = MultiprocessIterator(dataset['train'], config.train.batchsize)
+    test_iter = MultiprocessIterator(dataset['test'], config.train.batchsize, repeat=False, shuffle=False)
+    train_eval_iter = MultiprocessIterator(dataset['train_eval'], config.train.batchsize, repeat=False, shuffle=False)
+
+
+    # optimizer
+    def create_optimizer(model):
+        cp: Dict[str, Any] = copy(config.train.optimizer)
+        n = cp.pop('name').lower()
+
+        if n == 'adam':
+            optimizer = optimizers.Adam(**cp)
+        elif n == 'sgd':
+            optimizer = optimizers.SGD(**cp)
+        else:
+            raise ValueError(n)
+
+        optimizer.setup(model)
+        return optimizer
+
+
+    opts = {key: create_optimizer(model) for key, model in models.items()}
+
+    # updater
+    converter = partial(convert.concat_examples, padding=0)
+    updater = Updater(
+        loss_config=config.loss,
+        predictor=predictor,
+        discriminator=discriminator,
+        device=config.train.gpu,
+        iterator=train_iter,
+        optimizer=opts,
+        converter=converter,
+    )
+
+    # trainer
+    trigger_log = (config.train.log_iteration, 'iteration')
+    trigger_snapshot = (config.train.snapshot_iteration, 'iteration')
+    trigger_stop = (config.train.stop_iteration, 'iteration') if config.train.stop_iteration is not None else None
+
+    trainer = training.Trainer(updater, stop_trigger=trigger_stop, out=arguments.output)
+    tb_writer = SummaryWriter(Path(arguments.output))
+
+    ext = extensions.Evaluator(test_iter, models, converter, device=config.train.gpu, eval_func=updater.forward)
+    trainer.extend(ext, name='test', trigger=trigger_log)
+    ext = extensions.Evaluator(train_eval_iter, models, converter, device=config.train.gpu, eval_func=updater.forward)
+    trainer.extend(ext, name='train', trigger=trigger_log)
+
+    trainer.extend(extensions.dump_graph('predictor/loss'))
+
+    ext = extensions.snapshot_object(predictor, filename='predictor_{.updater.iteration}.npz')
+    trainer.extend(ext, trigger=trigger_snapshot)
+
+    trainer.extend(extensions.LogReport(trigger=trigger_log))
+    trainer.extend(TensorBoardReport(writer=tb_writer), trigger=trigger_log)
+
+    if trigger_stop is not None:
+        trainer.extend(extensions.ProgressBar(trigger_stop))
+
+    trainer.run()