def test_linear():
'''
This function does no input validation, it just returns the thing
that was passed in.
'''
xs = [1, 5, True, None, 'foo']
for x in xs:
assert(x == activations.linear(x))
def test_linear():
'''
This function does no input validation, it just returns the thing
that was passed in.
'''
xs = [1, 5, True, None, 'foo']
for x in xs:
assert (x == activations.linear(x))
def compute_similarity(self, repeated_context_vectors,
repeated_query_vectors):
element_wise_multiply = repeated_context_vectors * repeated_query_vectors
concatenated_tensor = K.concatenate([
repeated_context_vectors, repeated_query_vectors,
element_wise_multiply
],
axis=-1)
dot_product = K.squeeze(K.dot(concatenated_tensor, self.kernel),
axis=-1)
return linear(dot_product + self.bias)
def build_model():
model = models.Sequential()
model.add(
LSTM(100, activation=activations.relu(), input_shape=(timeSteps, 1)))
model.add(Dropout(1))
#model.add(layers.Dense(20, activation='relu'))
model.add(layers.Dense(1, activation=activations.linear()))
print("made it2")
model.compile(optimizer=tf.keras.optimizers.SGD(learning_rate=0.001),
loss=tf.keras.losses.mse,
metrics=['acc'])
return model
def InvertedRes(input, expansion):
'''
Args:
input: input tensor
expansion: expand filters size
Output:
output: output tensor
'''
#Pointwise Convolution
x = Conv2D(expansion*3,(1,1), padding='same')(input)
x = BatchNormalization()(x)
x = ReLU()(x)
#Depthwise Convolution
x = DepthwiseConv2D((3,3), padding='same')(x)
x = BatchNormalization()(x)
x = ReLU()(x)
#Pointwise Convolution
x = Conv2D(3,(1,1))(x)
x = BatchNormalization()(x)
x = linear(x)
x = Add()([x, input])
return x
def test_linear(self):
x = np.random.random((10, 5))
self.assertAllClose(x, activations.linear(x))
sns.set()
plt.plot(net, act)
plt.ylabel('act', fontsize=20)
plt.xlabel('net', fontsize=20)
plt.title(title, fontsize=20)
plt.savefig('./{0}.jpg'.format(title), dpi=300)
plt.close()
# create a numpy array - TensorFlow defaults to single precision floating point
netnp = np.linspace(-5.0, 5.0, 1000, dtype='float32')
# convert to a TensorFlow tensor
nettf = tf.convert_to_tensor(netnp)
# linear activation function
acttf = kact.linear(nettf)
# need to convert from TensorFlow tensors to numpy arrays before plotting
# eval() is called because TensorFlow tensors have no values until they are "run"
plt_act(nettf.eval(), acttf.eval(), 'linear activation function')
# relu activation function
acttf = kact.relu(nettf)
plt_act(nettf.eval(), acttf.eval(), 'rectified linear (relu)')
# sigmoid activation function
acttf = kact.sigmoid(nettf)
plt_act(nettf.eval(), acttf.eval(), 'sigmoid')
# hard sigmoid activation function
acttf = kact.hard_sigmoid(nettf)
plt_act(nettf.eval(), acttf.eval(), 'hard sigmoid')
def test_linear():
xs = [1, 5, True, None]
for x in xs:
assert (x == activations.linear(x))
def test_linear():
xs = [1, 5, True, None]
for x in xs:
assert(x == activations.linear(x))
def testLinearity():
testValues = [1, 5, True, None]
for x in testValues:
assert (x == activations.linear(x))
def call(self, inputs, training=False, mask=None):
x = nn.relu(self.fc1(inputs))
x = nn.relu(self.fc2(x))
x = linear(self.fc3(x))
return x
