def setup_arch_flags(cls):
super().setup_arch_flags()
flags.DEFINE_integer('max_epoch',
default=200,
help='number of max_epoch')
flags.DEFINE_multi_integer('milestones',
default=[
10, 20, 30, 40, 50, 60, 70, 80, 90, 100,
110, 120, 140, 160
],
help='miletones for lr_decay')
flags.DEFINE_float('gamma', default=0.85, help='gamma for lr_decay')
flags.DEFINE_string('device', default='cpu', help='cpu or cuda?')
flags.DEFINE_integer('printfreq',
default=5,
help='how many output for an epoch')
flags.DEFINE_integer('num_admm_innerloop',
default=2,
help='how many output for an epoch')
flags.DEFINE_integer('num_workers',
default=1,
help='how many output for an epoch')
flags.DEFINE_integer('batch_size',
default=1,
help='how many output for an epoch')
flags.DEFINE_boolean('vis_during_training',
default=False,
help='matplotlib plot image during training')
flags.DEFINE_boolean('group',
default=False,
help='group patient to perform 3D dice')
def setup_arch_flags(cls):
""" Setup the arch_hparams """
flags.DEFINE_float('weight_decay',
default=0,
help='decay of learning rate schedule')
flags.DEFINE_float('lr', default=0.001, help='learning rate')
flags.DEFINE_boolean('amsgrad', default=False, help='amsgrad')
flags.DEFINE_integer('optim_inner_loop_num',
default=5,
help='optim_inner_loop_num')
flags.DEFINE_string('arch', default='enet', help='arch_name')
flags.DEFINE_integer('num_classes', default=2, help='num of classes')
flags.DEFINE_string('method', default='admm_gc_size', help='arch_name')
flags.DEFINE_boolean('ignore_negative',
default=False,
help='ignore negative examples in the training')
def setup_arch_flags(cls):
super().setup_arch_flags()
flags.DEFINE_boolean(
'individual_size_constraint',
default=True,
help='Individual size constraint for each input image')
flags.DEFINE_float('eps', default=0.001, help='Individual size eps')
flags.DEFINE_integer(
'global_upbound',
default=2000,
help='global upper bound if individual_size_constraint is False')
flags.DEFINE_integer(
'global_lowbound',
default=20,
help='global lower bound if individual_size_constraint is False')
SizeConstraint.setup_arch_flag()
assert hparams['batch_size']==1,hparams['batch_size']
def run(argv):
del argv
hparams = flags.FLAGS.flag_values_dict()
check_consistance(hparams)
train_dataset, val_dataset = build_datasets(hparams)
arch_hparams = extract_from_big_dict(hparams, AdmmGCSize.arch_hparam_keys)
torchnet = get_arch(arch_hparams['arch'], arch_hparams)
admm = get_method(hparams['method'], torchnet, **hparams)
criterion = get_loss_fn(hparams['loss'])
trainer = ADMM_Trainer(admm, [train_dataset, val_dataset], criterion, hparams)
trainer.start_training()
if __name__ == '__main__':
torch.manual_seed(41)
flags.DEFINE_string('dataroot', default='cardiac', help='the name of the dataset')
flags.DEFINE_boolean('data_aug', default=False, help='data_augmentation')
flags.DEFINE_string('loss',default='partial_ce',help='loss used in admm loop')
flags.DEFINE_boolean('data_equ',default=False, help='data equalization')
# AdmmSize.setup_arch_flags()
# AdmmGCSize.setup_arch_flags()
# ADMM_size_inequality.setup_arch_flags()
ADMM_reg_size_inequality.setup_arch_flags()
ADMM_Trainer.setup_arch_flags()
app.run(run)