def test_monophone_config(basic_corpus_dir, basic_dict_path, temp_dir):
am_trainer = TrainableAligner(
corpus_directory=basic_corpus_dir,
dictionary_path=basic_dict_path,
temporary_directory=temp_dir,
)
config = MonophoneTrainer(identifier="mono", worker=am_trainer)
config.compute_calculated_properties()
assert config.realignment_iterations == [
0,
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
12,
14,
16,
18,
20,
23,
26,
29,
32,
35,
38,
]
am_trainer.cleanup()
def test_triphone_config(basic_corpus_dir, basic_dict_path, temp_dir):
am_trainer = TrainableAligner(
corpus_directory=basic_corpus_dir,
dictionary_path=basic_dict_path,
temporary_directory=temp_dir,
)
config = TriphoneTrainer(identifier="tri", worker=am_trainer)
config.compute_calculated_properties()
assert config.realignment_iterations == [10, 20, 30]
am_trainer.cleanup()
def test_load_mono_train(basic_corpus_dir, basic_dict_path, temp_dir,
mono_train_config_path):
params = TrainableAligner.parse_parameters(mono_train_config_path)
am_trainer = TrainableAligner(corpus_directory=basic_corpus_dir,
dictionary_path=basic_dict_path,
temporary_directory=temp_dir,
**params)
for t in am_trainer.training_configs.values():
assert not t.use_mp
assert t.use_energy
assert not am_trainer.use_mp
assert am_trainer.use_energy
am_trainer.cleanup()
def test_lda_mllt_config(basic_corpus_dir, basic_dict_path, temp_dir):
am_trainer = TrainableAligner(
corpus_directory=basic_corpus_dir,
dictionary_path=basic_dict_path,
temporary_directory=temp_dir,
)
assert am_trainer.beam == 10
assert am_trainer.retry_beam == 40
assert am_trainer.align_options["beam"] == 10
assert am_trainer.align_options["retry_beam"] == 40
config = LdaTrainer(identifier="lda", worker=am_trainer)
config.compute_calculated_properties()
assert config.mllt_iterations == [2, 4, 6, 12]
am_trainer.cleanup()
def test_load_basic_train(basic_corpus_dir, basic_dict_path, temp_dir,
basic_train_config_path):
params = TrainableAligner.parse_parameters(basic_train_config_path)
am_trainer = TrainableAligner(corpus_directory=basic_corpus_dir,
dictionary_path=basic_dict_path,
temporary_directory=temp_dir,
**params)
assert am_trainer.beam == 100
assert am_trainer.retry_beam == 400
assert am_trainer.align_options["beam"] == 100
assert am_trainer.align_options["retry_beam"] == 400
for trainer in am_trainer.training_configs.values():
assert trainer.beam == 100
assert trainer.retry_beam == 400
assert trainer.align_options["beam"] == 100
assert trainer.align_options["retry_beam"] == 400
am_trainer.cleanup()
def test_typing(basic_corpus_dir, basic_dict_path, temp_dir):
am_trainer = TrainableAligner(
corpus_directory=basic_corpus_dir,
dictionary_path=basic_dict_path,
temporary_directory=temp_dir,
)
trainer = SatTrainer(identifier="sat", worker=am_trainer)
assert type(trainer).__name__ == "SatTrainer"
assert isinstance(trainer, TrainerMixin)
assert isinstance(trainer, AlignMixin)
assert isinstance(trainer, MfaWorker)
assert isinstance(am_trainer, MfaWorker)
def test_load(basic_corpus_dir, basic_dict_path, temp_dir, config_directory):
path = os.path.join(config_directory, "basic_train_config.yaml")
params = TrainableAligner.parse_parameters(path)
am_trainer = TrainableAligner(corpus_directory=basic_corpus_dir,
dictionary_path=basic_dict_path,
temporary_directory=temp_dir,
**params)
assert len(am_trainer.training_configs) == 4
assert isinstance(am_trainer.training_configs["monophone"],
MonophoneTrainer)
assert isinstance(am_trainer.training_configs["triphone"], TriphoneTrainer)
assert isinstance(am_trainer.training_configs[am_trainer.final_identifier],
SatTrainer)
path = os.path.join(config_directory, "out_of_order_config.yaml")
with pytest.raises(ConfigError):
params = TrainableAligner.parse_parameters(path)
am_trainer.cleanup()
def train_acoustic_model(args: Namespace, unknown_args: Optional[List[str]] = None) -> None:
"""
Run the acoustic model training
Parameters
----------
args: :class:`~argparse.Namespace`
Command line arguments
unknown_args: list[str]
Optional arguments that will be passed to configuration objects
"""
trainer = TrainableAligner(
corpus_directory=args.corpus_directory,
dictionary_path=args.dictionary_path,
temporary_directory=args.temporary_directory,
**TrainableAligner.parse_parameters(args.config_path, args, unknown_args),
)
try:
trainer.train()
if args.output_model_path is not None:
trainer.export_model(args.output_model_path)
if args.output_directory is not None:
output_format = getattr(args, "output_format", None)
trainer.export_files(
args.output_directory,
output_format,
include_original_text=getattr(args, "include_original_text", False),
)
except Exception:
trainer.dirty = True
raise
finally:
trainer.cleanup()