def corr2d_multi_in(X, K):
# for x, k in zip(X, K):
# print(d2l.corr2d(x, k))
return nd.add_n(*[d2l.corr2d(x, k) for x, k in zip(X, K)])
def corr2d_multi_in(X, K):
# First, traverse along the 0th dimension (channel dimension) of X and K.
# Then, add them together by using * to turn the result list into a
# positional argument of the add_n function
return nd.add_n(*[d2l.corr2d(x, k) for x, k in zip(X, K)])
def forward(self, x):
return d2l.corr2d(x, self.weight.data()) + self.bias.data()
def convolution_test():
X = np.array([[0, 1, 2], [3, 4, 5], [6, 7, 8]])
K = np.array([[0, 1], [2, 3]])
corr = d2l.corr2d(X, K)
print("shape {}, value {}".format(corr.shape, corr))
K = np.array([[0, 1], [2, 3]])
corr = d2l.corr2d(X, K)
print("shape {}, value {}".format(corr.shape, corr))
class Conv2D(nn.Block):
def __init__(self, kernel_size, **kwargs):
super().__init__(**kwargs)
self.weight = self.params.get('weight', shape=kernel_size)
self.bias = self.params.get('bias', shape=(1, ))
def forward(self, x):
return d2l.corr2d(x, self.weight.data()) + self.bias.data()
def sample_black_white_image():
X = np.ones((6, 8))
X[:, 2:6] = 0
return X
# kernel
K = np.array([[1, -1]])
Y = d2l.corr2d(sample_black_white_image(), K)
print("shape {}, value {}".format(Y.shape, Y))
# transposed
Y_t = d2l.corr2d(sample_black_white_image().T, K)
print("transposed : shape {}, value {}".format(Y_t.shape, Y_t))