def __init__(self, num_class=10):
super(LeNet5, self).__init__()
self.num_class = num_class
self.conv1 = combined.Conv2d(
1, 6, kernel_size=5, batchnorm=True, activation='relu6')
self.conv2 = combined.Conv2d(6, 16, kernel_size=5, activation='relu')
self.fc1 = combined.Dense(16 * 5 * 5, 120, activation='relu')
self.fc2 = combined.Dense(120, 84, activation='relu')
self.fc3 = combined.Dense(84, self.num_class)
self.max_pool2d = nn.MaxPool2d(kernel_size=2, stride=2)
self.flattern = nn.Flatten()
def _conv_bn(in_channel, out_channel, ksize, stride=1):
"""Get a conv2d batchnorm and relu layer."""
return nn.SequentialCell([
combined.Conv2d(in_channel,
out_channel,
kernel_size=ksize,
stride=stride,
batchnorm=True)
])
def __init__(self, inp, oup, stride, expend_ratio):
super(InvertedResidual, self).__init__()
self.stride = stride
assert stride in [1, 2]
hidden_dim = int(inp * expend_ratio)
self.use_res_connect = self.stride == 1 and inp == oup
if expend_ratio == 1:
self.conv = nn.SequentialCell([
combined.Conv2d(hidden_dim,
hidden_dim,
3,
stride,
group=hidden_dim,
batchnorm=True,
activation='relu6'),
combined.Conv2d(hidden_dim, oup, 1, 1, batchnorm=True)
])
else:
self.conv = nn.SequentialCell([
combined.Conv2d(inp,
hidden_dim,
1,
1,
batchnorm=True,
activation='relu6'),
combined.Conv2d(hidden_dim,
hidden_dim,
3,
stride,
group=hidden_dim,
batchnorm=True,
activation='relu6'),
combined.Conv2d(hidden_dim, oup, 1, 1, batchnorm=True)
])
self.add = P.TensorAdd()
