def forward(self, x):
new_features = super(_DenseLayer, self).forward(x)
if self.drop_rate > 0:
new_features = F.dropout(new_features,
p=self.drop_rate,
training=self.training)
return torchgraph.cat([x, new_features], 1)
def forward(self, x):
branch3x3 = self.branch3x3(x)
branch3x3dbl = self.branch3x3dbl_1(x)
branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)
branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)
branch_pool = self.max_pool(x)
outputs = [branch3x3, branch3x3dbl, branch_pool]
return torchgraph.cat(outputs, 1)
def forward(self, x):
branch3x3 = self.branch3x3_1(x)
branch3x3 = self.branch3x3_2(branch3x3)
branch7x7x3 = self.branch7x7x3_1(x)
branch7x7x3 = self.branch7x7x3_2(branch7x7x3)
branch7x7x3 = self.branch7x7x3_3(branch7x7x3)
branch7x7x3 = self.branch7x7x3_4(branch7x7x3)
branch_pool = self.max_pool(x)
outputs = [branch3x3, branch7x7x3, branch_pool]
return torchgraph.cat(outputs, 1)
def forward(self, x, x_prev):
x_relu = self.relu(x_prev)
# path 1
x_path1 = self.path_1(x_relu)
# path 2
x_path2 = self.path_2(x_relu[:, :, 1:, 1:])
# final path
x_left = self.final_path_bn(torchgraph.cat([x_path1, x_path2], 1))
x_right = self.conv_1x1(x)
return NormalCellBranchCombine(self, x_left, x_right)
def forward(self, x_conv0, x_stem_0):
x_left = self.conv_1x1(x_stem_0)
x_relu = self.relu(x_conv0)
# path 1
x_path1 = self.path_1(x_relu)
# path 2
x_path2 = self.path_2(x_relu[:, :, 1:, 1:])
# final path
x_right = self.final_path_bn(torchgraph.cat([x_path1, x_path2], 1))
return ReductionCellBranchCombine(self, x_left, x_right)
def forward(self, x):
branch1x1 = self.branch1x1(x)
branch3x3 = self.branch3x3_1(x)
branch3x3 = [
self.branch3x3_2a(branch3x3),
self.branch3x3_2b(branch3x3),
]
branch3x3 = torchgraph.cat(branch3x3, 1)
branch3x3dbl = self.branch3x3dbl_1(x)
branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)
branch3x3dbl = [
self.branch3x3dbl_3a(branch3x3dbl),
self.branch3x3dbl_3b(branch3x3dbl),
]
branch3x3dbl = torchgraph.cat(branch3x3dbl, 1)
branch_pool = self.avg_pool(x)
branch_pool = self.branch_pool(branch_pool)
outputs = [branch1x1, branch3x3, branch3x3dbl, branch_pool]
return torchgraph.cat(outputs, 1)
def forward(self, x):
branch1x1 = self.branch1x1(x)
branch5x5 = self.branch5x5_1(x)
branch5x5 = self.branch5x5_2(branch5x5)
branch3x3dbl = self.branch3x3dbl_1(x)
branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)
branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)
branch_pool = self.avg_pool(x)
branch_pool = self.branch_pool(branch_pool)
outputs = [branch1x1, branch5x5, branch3x3dbl, branch_pool]
return torchgraph.cat(outputs, 1)
def forward(self, inputs, context, inference=False):
"""
Execute the decoder.
:param inputs: tensor with inputs to the decoder
:param context: state of encoder, encoder sequence lengths and hidden
state of decoder's LSTM layers
:param inference: if True stores and repackages hidden state
"""
self.inference = inference
enc_context, enc_len, hidden = context
hidden = self.init_hidden(hidden)
x = self.embedder(inputs)
x, h, attn, scores = self.att_rnn(x, hidden[0], enc_context, enc_len)
self.append_hidden(h)
x = self.dropout(x)
x = torchgraph.cat((x, attn), dim=2)
x, h = self.rnn_layers[0](x, hidden[1])
self.append_hidden(h)
for i in range(1, len(self.rnn_layers)):
residual = x
x = self.dropout(x)
x = torchgraph.cat((x, attn), dim=2)
x, h = self.rnn_layers[i](x, hidden[i + 1])
self.append_hidden(h)
x = x + residual
x = self.classifier(x)
hidden = self.package_hidden()
return x, scores, [enc_context, enc_len, hidden]
def forward(self, x):
branch1x1 = self.branch1x1(x)
branch7x7 = self.branch7x7_1(x)
branch7x7 = self.branch7x7_2(branch7x7)
branch7x7 = self.branch7x7_3(branch7x7)
branch7x7dbl = self.branch7x7dbl_1(x)
branch7x7dbl = self.branch7x7dbl_2(branch7x7dbl)
branch7x7dbl = self.branch7x7dbl_3(branch7x7dbl)
branch7x7dbl = self.branch7x7dbl_4(branch7x7dbl)
branch7x7dbl = self.branch7x7dbl_5(branch7x7dbl)
branch_pool = self.avg_pool(x)
branch_pool = self.branch_pool(branch_pool)
outputs = [branch1x1, branch7x7, branch7x7dbl, branch_pool]
return torchgraph.cat(outputs, 1)
def ReductionCellBranchCombine(cell, x_left, x_right):
x_comb_iter_0_left = cell.comb_iter_0_left(x_left)
x_comb_iter_0_right = cell.comb_iter_0_right(x_right)
x_comb_iter_0 = x_comb_iter_0_left + x_comb_iter_0_right
x_comb_iter_1_left = cell.comb_iter_1_left(x_left)
x_comb_iter_1_right = cell.comb_iter_1_right(x_right)
x_comb_iter_1 = x_comb_iter_1_left + x_comb_iter_1_right
x_comb_iter_2_left = cell.comb_iter_2_left(x_left)
x_comb_iter_2_right = cell.comb_iter_2_right(x_right)
x_comb_iter_2 = x_comb_iter_2_left + x_comb_iter_2_right
x_comb_iter_3_right = cell.comb_iter_3_right(x_comb_iter_0)
x_comb_iter_3 = x_comb_iter_3_right + x_comb_iter_1
x_comb_iter_4_left = cell.comb_iter_4_left(x_comb_iter_0)
x_comb_iter_4_right = cell.comb_iter_4_right(x_left)
x_comb_iter_4 = x_comb_iter_4_left + x_comb_iter_4_right
x_out = torchgraph.cat(
[x_comb_iter_1, x_comb_iter_2, x_comb_iter_3, x_comb_iter_4], 1)
return x_out
def forward(self, x):
x = self.squeeze_activation(self.squeeze(x))
return torchgraph.cat([
self.expand1x1_activation(self.expand1x1(x)),
self.expand3x3_activation(self.expand3x3(x))
], 1)