def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = global_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x,
tr_y,
te_x,
te_y,
net_func=cifar10_sequential_cbn6d_wd,
optimizer=tf.train.RMSPropOptimizer,
optimizer_args={
'decay': 0.9,
'epsilon': 1e-8
},
n_epochs=125,
batch_size=64,
lr_decay_func=DivideAtRatesWithDecay(start=0.001,
divide_by=2,
at=[0.6, 0.8],
max_steps=125,
decay=1e-6),
weight_decay=0.0001)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = channel_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x,
tr_y,
te_x,
te_y,
net_func=cifar10_inception_v3_wd,
optimizer=tf.train.RMSPropOptimizer,
optimizer_args={
'momentum': 0.9,
'decay': 0.9,
'epsilon': 1.0
},
n_epochs=100,
batch_size=32,
aux_loss_weight=0.3,
label_smoothing=0.1,
lr_decay_func=ExponentialDecayValue(start=0.045,
decay_rate=0.94,
decay_steps=2),
weight_decay=0.00004)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = global_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_basic(
tr_x, tr_y, te_x, te_y, net_func=cifar10_sequential_cbn6d)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = global_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_basic(
tr_x, tr_y, te_x, te_y, net_func=cifar10_resnet_bottleneck_20)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = global_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_basic(
tr_x, tr_y, te_x, te_y, net_func=cifar10_bn_inception_v1)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = global_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_basic(tr_x,
tr_y,
te_x,
te_y,
net_func=partial(
cifar10_se_resnext_29,
cardinality=8,
group_width=16))
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = global_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x, tr_y, te_x, te_y,
net_func = partial(cifar10_sequential_clrn5d3_wd, drop_rate = 0.3),
optimizer = tf.train.AdamOptimizer,
optimizer_args = None,
n_epochs = 100,
batch_size = 128,
lr_decay_func = FixValue(value = 1e-4),
weight_decay = 0.0001
)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = channel_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x,
tr_y,
te_x,
te_y,
net_func=cifar10_bn_inception_v1_wd,
optimizer=tf.train.MomentumOptimizer,
optimizer_args={'momentum': 0.9},
n_epochs=250,
batch_size=128,
lr_decay_func=ExponentialDecay(start=0.01, stop=0.0001, max_steps=200),
weight_decay=0.00004)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = pixel_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x, tr_y, te_x, te_y,
net_func = cifar10_mobilenet_v2_wd,
optimizer = tf.train.RMSPropOptimizer,
optimizer_args = {'decay': 0.9, 'momentum': 0.9},
n_epochs = 300,
batch_size = 96,
lr_decay_func = DecayValue(
start = 0.045,
decay_rate = 0.98),
weight_decay = 0.00004
)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = global_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x,
tr_y,
te_x,
te_y,
net_func=partial(cifar10_sequential_c3d_selu_drop, drop_rate=0.5),
n_epochs=50,
batch_size=128,
lr_decay_func=ExponentialDecay(start=0.001, stop=0.0001, max_steps=50),
optimizer=tf.train.AdamOptimizer,
weight_decay=None,
augmentation=False)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = pixel_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x,
tr_y,
te_x,
te_y,
net_func=cifar10_shufflenet_wd,
optimizer=tf.train.MomentumOptimizer,
optimizer_args={'momentum': 0.9},
n_epochs=300,
batch_size=128,
lr_decay_func=LinearDecay(start=0.5, stop=0.0, max_steps=300),
weight_decay=4e-5)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = pixel_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x, tr_y, te_x, te_y,
net_func = cifar10_sequential_allconvC_wd,
optimizer = tf.train.MomentumOptimizer,
optimizer_args = {'momentum': 0.9},
n_epochs = 350,
batch_size = 64,
lr_decay_func = DivideAt(
start = 0.05,
divide_by = 10,
at_steps = [200, 250, 300]),
weight_decay = 0.001
)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = pixel_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x, tr_y, te_x, te_y,
net_func = cifar10_resnet_bottleneck_20_wd,
optimizer = tf.train.MomentumOptimizer,
optimizer_args = {'momentum': 0.9},
n_epochs = 200,
batch_size = 128,
lr_decay_func = DivideAtRates(
start = 0.1,
divide_by = 10,
at = [0.5, 0.75],
max_steps = 200),
weight_decay = 0.0001
)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = channel_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x,
tr_y,
te_x,
te_y,
net_func=partial(cifar10_xception_wd, drop_rate=0.5),
optimizer=tf.train.MomentumOptimizer,
optimizer_args={'momentum': 0.9},
n_epochs=100,
batch_size=32,
lr_decay_func=ExponentialDecayValue(start=0.045,
decay_rate=0.94,
decay_steps=2),
weight_decay=0.00001)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = pixel_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x, tr_y, te_x, te_y,
net_func = partial(cifar10_resnext_29_wd, cardinality = 8, group_width = 16),
optimizer = tf.train.MomentumOptimizer,
optimizer_args = {'momentum': 0.9},
n_epochs = 300,
batch_size = 128,
lr_decay_func = DivideAtRates(
start = 0.1,
divide_by = 10,
at = [0.5, 0.75],
max_steps = 300),
weight_decay = 0.0005
)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = channel_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x, tr_y, te_x, te_y,
net_func = partial(cifar10_densenet_40_wd, drop_rate = 0.0),
optimizer = tf.train.MomentumOptimizer,
optimizer_args = {'momentum': 0.9},
n_epochs = 300,
batch_size = 64,
lr_decay_func = DivideAtRates(
start = 0.1,
divide_by = 10,
at = [0.5, 0.75],
max_steps = 300),
weight_decay = 0.0001
)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
def main():
tr_x, tr_y, te_x, te_y = load_cifar10_data()
tr_x, te_x = global_mean_std(tr_x, te_x)
mean_acc, max_acc, min_acc = eval_net_custom(
tr_x,
tr_y,
te_x,
te_y,
net_func=cifar10_inception_v3,
optimizer=tf.train.AdamOptimizer,
optimizer_args=None,
n_epochs=50,
batch_size=128,
aux_loss_weight=0.3,
label_smoothing=0.1,
lr_decay_func=ExponentialDecay(start=0.01, stop=0.001, max_steps=50),
weight_decay=None,
augmentation=False)
print("Mean accuracy: ", mean_acc)
print("Max accuracy: ", max_acc)
print("Min accuracy: ", min_acc)
