def test_learning_rate_value(self):
lr = -1.0
with pytest.raises(ValueError):
dr.exponential_decay_lr(lr, decay_rate, total_step, step_per_epoch, decay_epoch)
with pytest.raises(ValueError):
dr.polynomial_decay_lr(lr, end_learning_rate, total_step, step_per_epoch, decay_epoch, power)
def test_exponential_decay():
lr1 = dr.exponential_decay_lr(learning_rate, decay_rate, total_step,
step_per_epoch, decay_epoch)
assert len(lr1) == total_step
lr2 = dr.exponential_decay_lr(learning_rate, decay_rate, total_step,
step_per_epoch, decay_epoch, True)
assert len(lr2) == total_step
def test_total_step1(self):
total_step1 = 2.0
with pytest.raises(ValueError):
dr.exponential_decay_lr(learning_rate, decay_rate, total_step1, step_per_epoch, decay_epoch)
with pytest.raises(ValueError):
dr.cosine_decay_lr(min_lr, max_lr, total_step1, step_per_epoch, decay_epoch)
with pytest.raises(ValueError):
dr.polynomial_decay_lr(learning_rate, end_learning_rate, total_step1, step_per_epoch, decay_epoch, power)
def test_decay_epoch1(self):
decay_epoch1 = 'm'
with pytest.raises(TypeError):
dr.exponential_decay_lr(learning_rate, decay_rate, total_step,
step_per_epoch, decay_epoch1)
with pytest.raises(TypeError):
dr.cosine_decay_lr(min_lr, max_lr, total_step, step_per_epoch,
decay_epoch1)
with pytest.raises(TypeError):
dr.polynomial_decay_lr(learning_rate, end_learning_rate,
total_step, step_per_epoch, decay_epoch1,
power)
parser.add_argument('--device_id', type=int, default=0, help='device id of GPU. (Default: 0)')
args = parser.parse_args()
if args.device_target == "CPU":
args.dataset_sink_mode = False
context.set_context(mode=context.GRAPH_MODE, device_target=args.device_target, device_id=args.device_id)
network = Inceptionv3(cfg.num_classes)
net_loss = nn.SoftmaxCrossEntropyWithLogits(is_grad=False, sparse=True,
reduction="mean", smooth_factor=cfg.label_smoothing_eps)
ds_train = create_dataset(args.data_path, cfg.batch_size, cfg.epoch_size)
step_per_epoch = ds_train.get_dataset_size()
total_step = step_per_epoch * cfg.epoch_size
lr = exponential_decay_lr(learning_rate=cfg.lr_init,
decay_rate=cfg.lr_decay_rate, total_step=total_step,
step_per_epoch=step_per_epoch, decay_epoch=cfg.lr_decay_epoch)
net_opt = nn.RMSProp(network.trainable_params(), learning_rate=lr,
decay=cfg.rmsprop_decay, momentum=cfg.rmsprop_momentum,
epsilon=cfg.rmsprop_epsilon)
time_cb = TimeMonitor(data_size=ds_train.get_dataset_size())
config_ck = CheckpointConfig(save_checkpoint_steps=cfg.save_checkpoint_steps,
keep_checkpoint_max=cfg.keep_checkpoint_max)
ckpoint_cb = ModelCheckpoint(prefix="checkpoint_inceptionv3", config=config_ck)
# summary_cb = SummaryCollector(args.summary_path,
# collect_freq=1,
# keep_default_action=False,
# collect_specified_data={'collect_graph': True})
model = Model(network, net_loss, net_opt, metrics={"Accuracy": Accuracy()})
print("============== Starting Training ==============")
def test_decay_rate_value(self):
rate = -1.0
with pytest.raises(ValueError):
dr.exponential_decay_lr(learning_rate, rate, total_step,
step_per_epoch, decay_epoch)
def test_decay_rate_type(self):
rate = 'a'
with pytest.raises(TypeError):
dr.exponential_decay_lr(learning_rate, rate, total_step,
step_per_epoch, decay_epoch)
def test_is_stair(self):
is_stair = 1
with pytest.raises(TypeError):
dr.exponential_decay_lr(learning_rate, decay_rate, total_step,
step_per_epoch, decay_epoch, is_stair)
