def main(config_file: str):
if config_file is None:
config_file = os.path.join(dirs.CONFIG_DIR,
'train_isic_aleatoric.yaml')
context = ctx.TorchTrainContext('cuda')
context.load_from_config(config_file)
build_train = data.BuildData(build_dataset=isic.BuildIsicDataset())
build_valid = data.BuildData(build_dataset=isic.BuildIsicDataset())
if not hasattr(context.config.others, 'is_log_sigma'):
raise ValueError('"is_log_sigma" entry missing in configuration file')
is_log_sigma = context.config.others.is_log_sigma
train_steps = [TrainStepWithEval(loss.AleatoricLoss(is_log_sigma))]
train = trainloop.Train(train_steps, only_validate=False)
assemble = assembler.Subject2dAssembler()
validate = trainloop.ValidateSubject(
[AleatoricPredictStep(is_log_sigma=is_log_sigma)], [EvalStep()],
assemble,
convert_fn=None)
hook = hooks.ReducedComposeTrainLoopHook([
hooks.TensorboardXHook(),
hooks.ConsoleLogHook(print_subject_results=False),
hooks.SaveBestModelHook(),
hooks.SaveNLastModelHook(3)
])
train(context, build_train, build_valid, validate, hook=hook)
def main(config_file: str):
if config_file is None:
config_file = os.path.join(dirs.CONFIG_DIR,
'train_isic_auxiliary_segm.yaml')
context = ctx.TorchTrainContext('cuda')
context.load_from_config(config_file)
if not hasattr(context.config.others, 'prediction_dir'):
raise ValueError('"others.prediction_dir" is required in the config')
prediction_dir = context.config.others.prediction_dir
build_train = data.BuildData(build_dataset=isic.BuildIsicDataset(),
prediction_dir=prediction_dir)
build_valid = data.BuildData(build_dataset=isic.BuildIsicDataset(),
prediction_dir=prediction_dir)
train_steps = [SpecialTrainStep(), step.EvalStep()]
train = trainloop.Train(train_steps, only_validate=False)
assemble = assembler.Subject2dAssembler()
validate = trainloop.ValidateSubject([SpecialSegmentationPredictStep()],
[EvalSubjectStep()],
assemble,
entries=('probabilities', 'labels'))
hook = hooks.ReducedComposeTrainLoopHook([
hooks.TensorboardXHook(),
hooks.ConsoleLogHook(print_subject_results=False),
hooks.SaveBestModelHook(),
hooks.SaveNLastModelHook(3)
])
train(context, build_train, build_valid, validate, hook=hook)
def main(config_file):
if config_file is None:
config_file = os.path.join(dirs.CONFIG_DIR, 'test_isic_aleatoric.yaml')
context = ctx.TorchTestContext('cuda')
context.load_from_config(config_file)
build_test = data.BuildData(build_dataset=isic.BuildIsicDataset(), )
if not hasattr(context.config.others, 'is_log_sigma'):
raise ValueError('"is_log_sigma" entry missing in configuration file')
is_log_sigma = context.config.others.is_log_sigma
test_steps = [AleatoricPredictStep(is_log_sigma), PrepareSubjectStep()]
subject_steps = [EvalSubjectStep()]
subject_assembler = assembler.Subject2dAssembler()
test = loop.Test(test_steps, subject_steps, subject_assembler)
hook = hooks.ReducedComposeTestLoopHook([
hooks.ConsoleTestLogHook(),
hooks.WriteTestMetricsCsvHook('metrics.csv'),
WriteHook()
])
test(context, build_test, hook=hook)
def main(config_file):
if config_file is None:
config_file = os.path.join(dirs.CONFIG_DIR,
'test_isic_auxiliary_segm.yaml')
context = ctx.TorchTestContext('cuda')
context.load_from_config(config_file)
if not hasattr(context.config.others, 'prediction_dir'):
raise ValueError('"others.prediction_dir" is required in the config')
prediction_dir = context.config.others.prediction_dir
build_test = data.BuildData(build_dataset=isic.BuildIsicDataset(),
prediction_dir=prediction_dir)
test_steps = [SegmentationPredictStep()]
subject_steps = [EvalSubjectStep()]
subject_assembler = assembler.Subject2dAssembler()
test = loop.Test(test_steps,
subject_steps,
subject_assembler,
entries=('probabilities', 'labels', 'orig_prediction'))
hook = hooks.ReducedComposeTestLoopHook([
hooks.ConsoleTestLogHook(),
hooks.WriteTestMetricsCsvHook('metrics.csv'),
WriteHook()
])
test(context, build_test, hook=hook)
def main(config_file: str, config_id: str):
if config_file is None:
if config_id == 'baseline':
config_file = os.path.join(dirs.CONFIG_DIR,
'train_isic_baseline.yaml')
elif config_id == 'center':
config_file = os.path.join(dirs.CONFIG_DIR,
'train_isic_center.yaml')
elif config_id in ('cv0', 'cv1', 'cv2', 'cv3', 'cv4'):
config_file = os.path.join(
dirs.CONFIG_DIR, 'baseline_cv',
'train_isic_baseline_cv{}.yaml'.format(config_id[-1]))
elif config_id in ('ensemble0', 'ensemble1', 'ensemble2', 'ensemble3',
'ensemble4', 'ensemble5', 'ensemble6', 'ensemble7',
'ensemble8', 'ensemble9'):
config_file = os.path.join(
dirs.CONFIG_DIR, 'train_ensemble',
'train_isic_ensemble_{}.yaml'.format(config_id[-1]))
else:
config_file = os.path.join(dirs.CONFIG_DIR,
'train_isic_baseline.yaml')
context = ctx.TorchTrainContext('cuda')
context.load_from_config(config_file)
build_train = data.BuildData(build_dataset=isic.BuildIsicDataset())
build_valid = data.BuildData(build_dataset=isic.BuildIsicDataset())
train_steps = [step.TrainStep(), step.EvalStep()]
train = trainloop.Train(train_steps, only_validate=False)
assemble = assembler.Subject2dAssembler()
validate = trainloop.ValidateSubject(
[step.SegmentationPredictStep(has_labels=True),
PrepareSubjectStep()], [EvalStep()],
assemble,
convert_fn=None)
hook = hooks.ReducedComposeTrainLoopHook([
hooks.TensorboardXHook(),
hooks.ConsoleLogHook(print_subject_results=False),
hooks.SaveBestModelHook(),
hooks.SaveNLastModelHook(3)
])
train(context, build_train, build_valid, validate, hook=hook)
def main(config_file: str):
if config_file is None:
config_file = os.path.join(dirs.CONFIG_DIR,
'train_isic_auxiliary_feat.yaml')
context = ctx.TorchTrainContext('cuda')
context.load_from_config(config_file)
if hasattr(context.config.others, 'model_dir') and hasattr(
context.config.others, 'test_at'):
mf = mgt.ModelFiles.from_model_dir(context.config.others.model_dir)
checkpoint_path = mgt.model_service.find_checkpoint_file(
mf.weight_checkpoint_dir, context.config.others.test_at)
model = mgt.model_service.load_model_from_parameters(
mf.model_path(), with_optimizer=False)
model.provide_features = True
mgt.model_service.load_checkpoint(checkpoint_path, model)
test_model = model.to(context.device)
test_model.eval()
for params in test_model.parameters():
params.requires_grad = False
build_train = data.BuildData(build_dataset=isic.BuildIsicDataset())
build_valid = data.BuildData(build_dataset=isic.BuildIsicDataset())
train_steps = [SpecialTrainStep(test_model), step.EvalStep()]
train = trainloop.Train(train_steps, only_validate=False)
assemble = assembler.Subject2dAssembler()
validate = trainloop.ValidateSubject(
[SpecialSegmentationPredictStep(test_model),
PrepareSubjectStep()], [EvalSubjectStep()],
assemble,
entries=('probabilities', 'net_predictions', 'labels'))
hook = hooks.ReducedComposeTrainLoopHook([
hooks.TensorboardXHook(),
hooks.ConsoleLogHook(print_subject_results=False),
hooks.SaveBestModelHook(),
hooks.SaveNLastModelHook(3)
])
train(context, build_train, build_valid, validate, hook=hook)
def main(config_file):
if config_file is None:
config_file = os.path.join(dirs.CONFIG_DIR, 'test_isic_ensemble.yaml')
context = ctx.TorchTestContext('cuda')
context.load_from_config(config_file)
build_test = data.BuildData(build_dataset=isic.BuildIsicDataset(), )
if not hasattr(context.config.others, 'model_dir') or not hasattr(
context.config.others, 'test_at'):
raise ValueError(
'missing "model_dir" or "test_at" entry in the configuration (others)'
)
model_dirs = context.config.others.model_dir
if isinstance(model_dirs, str):
model_dirs = [model_dirs]
test_models = []
for i, model_dir in enumerate(model_dirs):
logging.info('load additional model [{}/{}] {}'.format(
i + 1, len(model_dirs), os.path.basename(model_dir)))
mf = mgt.ModelFiles.from_model_dir(model_dir)
checkpoint_path = mgt.model_service.find_checkpoint_file(
mf.weight_checkpoint_dir, context.config.others.test_at)
model = mgt.model_service.load_model_from_parameters(
mf.model_path(), with_optimizer=False)
mgt.model_service.load_checkpoint(checkpoint_path, model)
test_model = model.to(context.device)
test_model.eval()
for params in test_model.parameters():
params.requires_grad = False
test_models.append(test_model)
test_steps = [
EnsemblePredictionStep(test_models),
customstep.MultiPredictionSummary(),
PrepareSubjectStep()
]
subject_steps = [EvalSubjectStep()]
subject_assembler = assembler.Subject2dAssembler()
test = loop.Test(test_steps,
subject_steps,
subject_assembler,
entries=None)
hook = hooks.ReducedComposeTestLoopHook([
hooks.ConsoleTestLogHook(),
hooks.WriteTestMetricsCsvHook('metrics.csv'),
WriteHook()
])
test(context, build_test, hook=hook)
def main(config_file):
if config_file is None:
config_file = os.path.join(dirs.CONFIG_DIR,
'test_isic_auxiliary_feat.yaml')
context = ctx.TorchTestContext('cuda')
context.load_from_config(config_file)
build_test = data.BuildData(build_dataset=isic.BuildIsicDataset(), )
if hasattr(context.config.others, 'model_dir') and hasattr(
context.config.others, 'test_at'):
mf = mgt.ModelFiles.from_model_dir(context.config.others.model_dir)
checkpoint_path = mgt.model_service.find_checkpoint_file(
mf.weight_checkpoint_dir, context.config.others.test_at)
model = mgt.model_service.load_model_from_parameters(
mf.model_path(), with_optimizer=False)
model.provide_features = True
mgt.model_service.load_checkpoint(checkpoint_path, model)
test_model = model.to(context.device)
test_model.eval()
for params in test_model.parameters():
params.requires_grad = False
test_steps = [SegmentationPredictStep(test_model), PrepareSubjectStep()]
subject_steps = [EvalSubjectStep()]
subject_assembler = assembler.Subject2dAssembler()
test = loop.Test(test_steps,
subject_steps,
subject_assembler,
entries=('probabilities', 'segm_probabilities', 'labels'))
hook = hooks.ReducedComposeTestLoopHook([
hooks.ConsoleTestLogHook(),
hooks.WriteTestMetricsCsvHook('metrics.csv'),
WriteHook()
])
test(context, build_test, hook=hook)
def main(config_file, config_id):
if config_file is None:
if config_id == 'baseline':
config_file = os.path.join(dirs.CONFIG_DIR, 'test_isic_baseline.yaml')
elif config_id == 'baseline_mc':
config_file = os.path.join(dirs.CONFIG_DIR, 'test_isic_baseline_mc.yaml')
elif config_id == 'center':
config_file = os.path.join(dirs.CONFIG_DIR, 'test_isic_center.yaml')
elif config_id == 'center_mc':
config_file = os.path.join(dirs.CONFIG_DIR, 'test_isic_center_mc.yaml')
elif config_id in ('cv0', 'cv1', 'cv2', 'cv3', 'cv4'):
config_file = os.path.join(dirs.CONFIG_DIR, 'baseline_cv',
'test_isic_baseline_cv{}.yaml'.format(config_id[-1]))
else:
config_file = os.path.join(dirs.CONFIG_DIR, 'test_isic_baseline.yaml')
context = ctx.TorchTestContext('cuda')
context.load_from_config(config_file)
build_test = data.BuildData(
build_dataset=isic.BuildIsicDataset(),
)
if hasattr(context.config.others, 'mc'):
test_steps = [customstep.McPredictStep(context.config.others.mc),
customstep.MultiPredictionSummary()]
else:
test_steps = [step.SegmentationPredictStep(do_probs=True)]
test_steps.append(PrepareSubjectStep())
subject_steps = [EvalSubjectStep()]
subject_assembler = assembler.Subject2dAssembler()
test = loop.Test(test_steps, subject_steps, subject_assembler)
hook = hooks.ReducedComposeTestLoopHook([hooks.ConsoleTestLogHook(),
hooks.WriteTestMetricsCsvHook('metrics.csv'),
WriteHook()
])
test(context, build_test, hook=hook)