def test_dropout_dynamic_same_result():
x = mge.ones(10)
R.manual_seed(0)
a = F.dropout(x, 0.5)
R.manual_seed(0)
b = F.dropout(x, 0.5)
assert np.all(a.numpy() == b.numpy())
def forward(self, x):
new_features = self.feature(x)
if self.drop_rate > 0:
new_features = F.dropout(new_features,
drop_prob=self.drop_rate,
rescale=self.training)
return F.concat([x, new_features], 1)
def forward(self, x):
# aux1: N x 512 x 14 x 14, aux2: N x 528 x 14 x 14
x = self.avgpool(x)
# aux1: N x 512 x 4 x 4, aux2: N x 528 x 4 x 4
x = self.conv(x)
# N x 128 x 4 x 4
x = x.reshape(x.shape[0], -1)
# N x 2048
x = F.relu(self.fc1(x))
# N x 1024
x = F.dropout(x, 0.7)
# N x 1024
x = self.fc2(x)
# N x num_classes
return x
def test_dropout():
data = tensor(np.ones(10, dtype=np.float32))
out = F.dropout(data, 1.0 / 3.0, training=False)
assert out.numpy().sum() >= 0.0
def forward(self, inps):
return F.dropout(inps[0], self.param["ratio"], self.param["training"])
def test_dropout_dynamic_diff_result():
x = mge.ones(10)
a = F.dropout(x, 0.5)
b = F.dropout(x, 0.5)
assert np.any(a.numpy() != b.numpy())
[(32, 32, 16, 16, 16)],
True,
1000,
),
(
"convTranspose2d",
lambda x: module_cache["ConvTranspose2d"][0](x),
lambda x: module_cache["ConvTranspose2d"][1](x),
[(2, 32, 16, 16)],
[(32, 32, 128, 128)],
True,
1000,
),
(
"dropout",
lambda x: MF.dropout(x, 0.5),
TF.dropout,
[(100, 100)],
[(64, 512, 16, 16)],
True,
1000,
),
(
"dw_conv2d",
lambda x: module_cache["dw_conv2d"][0](x),
lambda x: module_cache["dw_conv2d"][1](x),
[(2, 32, 16, 16)],
[(32, 32, 128, 128)],
True,
1000,
),
