# dataset
train_data = gluon.data.DataLoader(dataset=IndianDataset(train=True),
batch_size=batch_size,
shuffle=True,
last_batch='rollover')
val_data = gluon.data.DataLoader(dataset=IndianDataset(train=False),
batch_size=batch_size,
shuffle=False)
# g(x) : stochastic input augmentation function
#def g(x):
# return x + nd.random.normal(0,stochastic_ratio,shape=x.shape)
# model
basemodel_zoo = 'simple2'
net = symbols.get_model(basemodel_zoo)
#net_t = symbols.get_model(basemodel_zoo)
net.initialize(mx.init.Xavier(magnitude=2.24))
#net.initialize(mx.init.MSRAPrelu())
#net_t.initialize(mx.init.MSRAPrelu())
#net.initialize(mx.init.Normal(0.5) ,ctx=ctx)
#net.load_parameters(para_filepath)
#net_t.load_parameters(para_filepath)
net.collect_params().reset_ctx(ctx)
#net_t.collect_params().reset_ctx(ctx)
# solve
loss = gloss.SoftmaxCrossEntropyLoss()
metric = mx.metric.Accuracy()
transform = lambda data, label: (data.reshape(784, ).astype(np.float32) / 255,
label)
train_data = gluon.data.DataLoader(dataset=gluon.data.vision.MNIST(
train=True, transform=transform),
batch_size=100,
shuffle=True,
last_batch='discard')
val_data = gluon.data.DataLoader(dataset=gluon.data.vision.MNIST(
train=False, transform=transform),
batch_size=100,
shuffle=False)
# network
modelname = 'semi_pi_simple2'
basemodel_zoo = 'simple2'
net = symbols.get_model('simple2')
net.initialize(mx.init.Xavier(magnitude=2.24))
#net.load_parameters(os.path.join('symbols','para','%s.params'%(modelname)))
# g(x) : stochastic input augmentation function
def g(x):
return x + nd.random.normal(0, stochastic_ratio, shape=x.shape)
# loss function
l_logistic = gloss.SoftmaxCrossEntropyLoss()
l_l2loss = gloss.L2Loss()
metric = mx.metric.Accuracy()
batch_size=batch_size,
shuffle=True,
last_batch='rollover')
val_data = gluon.data.DataLoader(dataset=IndianDataset(train=False),
batch_size=batch_size,
shuffle=False)
# g(x) : stochastic input augmentation function
def g(x):
return x #+ nd.random.normal(0,stochastic_ratio,shape=x.shape)
# model
basemodel_zoo = 'simple2'
net = symbols.get_model(basemodel_zoo, pretrained=True)
net_t = symbols.get_model(basemodel_zoo, pretrained=True)
#net.initialize(mx.init.Xavier(magnitude=2.24))
#net.initialize(mx.init.MSRAPrelu())
#net_t.initialize(mx.init.MSRAPrelu())
#net.initialize(mx.init.Normal(0.5) ,ctx=ctx)
#net.load_parameters(para_filepath)
#net_t.load_parameters(para_filepath)
net.collect_params().reset_ctx(ctx)
net_t.collect_params().reset_ctx(ctx)
# solve
l_logistic = gloss.SoftmaxCrossEntropyLoss()
l_l2loss = gloss.L2Loss()
metric = mx.metric.Accuracy()