本文整理汇总了Python中tensorflow.sign函数的典型用法代码示例。如果您正苦于以下问题:Python sign函数的具体用法?Python sign怎么用?Python sign使用的例子?那么恭喜您, 这里精选的函数代码示例或许可以为您提供帮助。
在下文中一共展示了sign函数的20个代码示例,这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞,您的评价将有助于我们的系统推荐出更棒的Python代码示例。
示例1: ternary_operation
def ternary_operation(x):
"""Ternary operation use threshold computed with weights."""
g = tf.compat.v1.get_default_graph()
with g.gradient_override_map({"Sign": "Identity"}):
threshold = _compute_threshold(x)
x = tf.sign(tf.add(tf.sign(tf.add(x, threshold)), tf.sign(tf.add(x, -threshold))))
return x
开发者ID:zsdonghao,项目名称:tensorlayer,代码行数:7,代码来源:utils.py
示例2: noisy_activation
def noisy_activation(x, generic, linearized, training, alpha=1.1, c=0.5):
"""
Implements the noisy activation with Half-Normal Noise for Hard-Saturation
functions. See http://arxiv.org/abs/1603.00391, Algorithm 1.
Args:
x: Tensor which is an input to the activation function
generic: The generic formulation of the activation function. (denoted
as h in the paper)
linearized: Linearization of the activation based on the first-order
Tailor expansion around zero. (denoted as u in the paper)
training: A boolean tensor telling whether we are in the training stage
(and the noise is sampled) or in runtime when the expactation is
used instead.
alpha: Mixing hyper-parameter. The leakage rate from the linearized
function to the nonlinear one.
c: Standard deviation of the sampled noise.
"""
delta = generic(x) - linearized(x)
d = -tf.sign(x) * tf.sign(1 - alpha)
p = tf.Variable(1.0)
scale = c * (tf.sigmoid(p * delta) - 0.5) ** 2
noise = tf.select(training, tf.abs(tf.random_normal([])), math.sqrt(2 / math.pi))
activation = alpha * generic(x) + (1 - alpha) * linearized(x) + d * scale * noise
return activation
开发者ID:alvaz16,项目名称:neuralmonkey,代码行数:33,代码来源:noisy_gru_cell.py
示例3: get_accuracy_loss
def get_accuracy_loss(arg,x,y,y_):
'''
Note: when the task is regression accuracy = loss but for classification
loss = cross_entropy,svm_loss, surrogate_loss, etc and accuracy = 1 - {0-1 loss}.
'''
with tf.name_scope("loss_and_acc") as scope:
# loss
if arg.softmax:
#cross_entropy = tf.reduce_mean(-tf.rduce_sum(y_ * tf.log(y), reduction_indices=[1]))
diff = tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y)
cross_entropy = tf.reduce_mean(diff)
loss = cross_entropy
correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(y_,1)) # list of booleans indicating correct predictions
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
else:
l2_loss = tf.reduce_sum( tf.reduce_mean(tf.square(y_-y), 0))
loss = l2_loss
y = tf.cast(tf.sign(y),tf.float32)
y_ = tf.cast(tf.sign(y_),tf.float32)
correct_prediction = tf.equal(y, y_) # list of booleans indicating correct predictions
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
# accuracy
# if arg.classification:
# correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(y_,1)) # list of booleans indicating correct predictions
# accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
# else:
# accuracy = tf.equal(tf.argmax(y,1), tf.argmax(y_,1))
##
tf.summary.scalar('loss', loss)
tf.summary.scalar('accuracy', accuracy)
return loss, accuracy
开发者ID:brando90,项目名称:hbf_tensorflow_code,代码行数:31,代码来源:main_hp.py
示例4: loss_func
def loss_func(score_op):
final_scores = tf.placeholder(tf.float32, shape=[None])
squared_errors = tf.square(tf.reshape(score_op, [-1]) - final_scores)
#mean_sq_err = tf.reduce_mean(squared_errors, name='mean_sq_err')
cross_entropy_ish_loss = tf.reduce_mean(-tf.log(tf.constant(1.0) - tf.constant(0.5) * tf.abs(tf.reshape(score_op, [-1]) - final_scores), name='cross-entropy-ish-loss'))
correct_prediction = tf.equal(tf.sign(tf.reshape(score_op, [-1])), tf.sign(final_scores))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32), name='accuracy')
#return final_scores, mean_sq_err, accuracy, squared_errors
return final_scores, cross_entropy_ish_loss, accuracy
开发者ID:TheDuck314,项目名称:go-NN,代码行数:11,代码来源:EvalTraining.py
示例5: loss_func
def loss_func(logits):
final_maps = tf.placeholder(tf.float32, shape=[None, 361])
# final maps are originally -1 to 1. rescale them to 0 to 1 probabilities:
final_prob_maps = final_maps * tf.constant(0.5) + tf.constant(0.5)
cross_entropy = tf.nn.sigmoid_cross_entropy_with_logits(logits=logits, targets=final_prob_maps)
cross_entropy_mean = tf.reduce_mean(cross_entropy, name='cross_entropy_mean')
correct_prediction = tf.equal(tf.sign(logits), tf.sign(final_maps))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
return final_maps, cross_entropy_mean, accuracy
开发者ID:TheDuck314,项目名称:go-NN,代码行数:11,代码来源:InfluenceTraining.py
示例6: angular_symmetry
def angular_symmetry(self, d_cutoff, d, atom_numbers, coordinates):
""" Angular Symmetry Function """
max_atoms = self.max_atoms
embedding = tf.eye(np.max(self.atom_cases) + 1)
atom_numbers_embedded = tf.nn.embedding_lookup(embedding, atom_numbers)
Rs = np.linspace(0., self.angular_cutoff, self.angular_length)
ita = 3 / (Rs[1] - Rs[0])**2
thetas = np.linspace(0., np.pi, self.angular_length)
zeta = float(self.angular_length**2)
ita, zeta, Rs, thetas = np.meshgrid(ita, zeta, Rs, thetas)
zeta = tf.cast(np.reshape(zeta, (1, 1, 1, 1, -1)), tf.float32)
ita = tf.cast(np.reshape(ita, (1, 1, 1, 1, -1)), tf.float32)
Rs = tf.cast(np.reshape(Rs, (1, 1, 1, 1, -1)), tf.float32)
thetas = tf.cast(np.reshape(thetas, (1, 1, 1, 1, -1)), tf.float32)
length = zeta.get_shape().as_list()[-1]
vector_distances = tf.stack([coordinates] * max_atoms, 1) - tf.stack(
[coordinates] * max_atoms, 2)
R_ij = tf.stack([d] * max_atoms, axis=3)
R_ik = tf.stack([d] * max_atoms, axis=2)
f_R_ij = tf.stack([d_cutoff] * max_atoms, axis=3)
f_R_ik = tf.stack([d_cutoff] * max_atoms, axis=2)
# Define angle theta = arccos(R_ij(Vector) dot R_ik(Vector)/R_ij(distance)/R_ik(distance))
vector_mul = tf.reduce_sum(tf.stack([vector_distances] * max_atoms, axis=3) * \
tf.stack([vector_distances] * max_atoms, axis=2), axis=4)
vector_mul = vector_mul * tf.sign(f_R_ij) * tf.sign(f_R_ik)
theta = tf.acos(tf.math.divide(vector_mul, R_ij * R_ik + 1e-5))
R_ij = tf.stack([R_ij] * length, axis=4)
R_ik = tf.stack([R_ik] * length, axis=4)
f_R_ij = tf.stack([f_R_ij] * length, axis=4)
f_R_ik = tf.stack([f_R_ik] * length, axis=4)
theta = tf.stack([theta] * length, axis=4)
out_tensor = tf.pow((1. + tf.cos(theta - thetas)) / 2., zeta) * \
tf.exp(-ita * tf.square((R_ij + R_ik) / 2. - Rs)) * f_R_ij * f_R_ik * 2
if self.atomic_number_differentiated:
out_tensors = []
for id_j, atom_type_j in enumerate(self.atom_cases):
for atom_type_k in self.atom_cases[id_j:]:
selected_atoms = tf.stack([atom_numbers_embedded[:, :, atom_type_j]] * max_atoms, axis=2) * \
tf.stack([atom_numbers_embedded[:, :, atom_type_k]] * max_atoms, axis=1)
selected_atoms = tf.expand_dims(
tf.expand_dims(selected_atoms, axis=1), axis=4)
out_tensors.append(
tf.reduce_sum(out_tensor * selected_atoms, axis=(2, 3)))
return tf.concat(out_tensors, axis=2)
else:
return tf.reduce_sum(out_tensor, axis=(2, 3))
开发者ID:ktaneishi,项目名称:deepchem,代码行数:54,代码来源:transformers.py
示例7: main
def main():
data, labels = input()
# logits=inference()
# loss_step=loss(logits)
# train_step = train(loss_step,0.001)
# sess=tf.Session()
# sess.run(tf.global_variables_initializer())
# print(sess.run([w,b]))
# labels.shape=(6000,1)
# for i in range(1000):
# sess.run(train_step,feed_dict={x_placehold:data,y_placehold:labels})
# wc,bc=sess.run([w,b],feed_dict={x_placehold:data,y_placehold:labels})
# print(wc,bc)
labels.shape = (6000, 1)
print(data)
print(labels)
print(data.shape)
print(labels.shape)
print(feed_dict('D'))
sess = tf.Session()
sess.run(tf.global_variables_initializer())
# sess.run(tf.sign(tf.matmul(x_placehold, w) + b),
# feed_dict=feed_dict('D'))
sess.run(tf.sign(tf.matmul(x_placehold, w) + b) - y_placehold,
feed_dict=feed_dict())
sess.run(tf.square(tf.sign(tf.matmul(x_placehold, w) + b) -
y_placehold), feed_dict=feed_dict())
sess.run(tf.reduce_sum(tf.square(tf.sign(tf.matmul(x_placehold, w) + b) -
y_placehold)), feed_dict=feed_dict())
logits = tf.matmul(x_placehold, w) + b
loss_op = tf.reduce_sum(tf.square(logits - y_placehold))
with tf.name_scope('loss'):
tf.summary.scalar('error', loss_op)
with tf.name_scope('w'):
tf.summary.scalar('x', w[0, 0])
tf.summary.scalar('y', w[1, 0])
merged = tf.summary.merge_all()
train_writer = tf.summary.FileWriter('./train', sess.graph)
optimizer = tf.train.GradientDescentOptimizer(0.1)
for i in range(1000):
sess.run(optimizer.minimize(loss_op), feed_dict=feed_dict())
summary = sess.run(merged, feed_dict=feed_dict())
train_writer.add_summary(summary, i)
wc, bc = sess.run([w, b], feed_dict=feed_dict())
print(wc, bc)
开发者ID:weitaoli-echuang,项目名称:neural-networks-and-deep-learning,代码行数:53,代码来源:1.py
示例8: neural_attention
def neural_attention(embedding_dim=384, encoding_dim=128):
embeddings = tf.Variable(tf.random_normal([vocab_size, embedding_dim], stddev=0.22), dtype=tf.float32)
tf.contrib.layers.apply_regularization(tf.contrib.layers.l2_regularizer(1e-4), [embeddings])
with tf.variable_scope('encode'):
with tf.variable_scope('X'):
X_lens = tf.reduce_sum(tf.sign(tf.abs(X)), 1)
embedded_X = tf.nn.embedding_lookup(embeddings, X)
encoded_X = tf.nn.dropout(embedded_X, keep_prob)
gru_cell = tf.contrib.rnn.core_rnn_cell.GRUCell(encoding_dim)
outputs, output_states = tf.nn.bidirectional_dynamic_rnn(gru_cell, gru_cell, encoded_X,
sequence_length=X_lens, dtype=tf.float32,
swap_memory=True)
encoded_X = tf.concat(outputs, 2)
with tf.variable_scope('Q'):
Q_lens = tf.reduce_sum(tf.sign(tf.abs(Q)), 1)
embedded_Q = tf.nn.embedding_lookup(embeddings, Q)
encoded_Q = tf.nn.dropout(embedded_Q, keep_prob)
gru_cell = tf.contrib.rnn.core_rnn_cell.GRUCell(encoding_dim)
outputs, output_states = tf.nn.bidirectional_dynamic_rnn(gru_cell, gru_cell, encoded_Q,
sequence_length=Q_lens, dtype=tf.float32,
swap_memory=True)
encoded_Q = tf.concat(outputs, 2)
W_q = tf.Variable(tf.random_normal([2 * encoding_dim, 4 * encoding_dim], stddev=0.22), dtype=tf.float32)
b_q = tf.Variable(tf.random_normal([2 * encoding_dim, 1], stddev=0.22), dtype=tf.float32)
W_d = tf.Variable(tf.random_normal([2 * encoding_dim, 6 * encoding_dim], stddev=0.22), dtype=tf.float32)
b_d = tf.Variable(tf.random_normal([2 * encoding_dim, 1], stddev=0.22), dtype=tf.float32)
g_q = tf.Variable(tf.random_normal([10 * encoding_dim, 2 * encoding_dim], stddev=0.22), dtype=tf.float32)
g_d = tf.Variable(tf.random_normal([10 * encoding_dim, 2 * encoding_dim], stddev=0.22), dtype=tf.float32)
with tf.variable_scope('attend') as scope:
infer_gru = tf.contrib.rnn.core_rnn_cell.GRUCell(4 * encoding_dim)
infer_state = infer_gru.zero_state(batch_size, tf.float32)
for iter_step in range(8):
if iter_step > 0:
scope.reuse_variables()
_, q_glimpse = glimpse(W_q, b_q, encoded_Q, infer_state)
d_attention, d_glimpse = glimpse(W_d, b_d, encoded_X, tf.concat([infer_state, q_glimpse], 1 ))
gate_concat = tf.concat([infer_state, q_glimpse, d_glimpse, q_glimpse * d_glimpse], 1)
r_d = tf.sigmoid(tf.matmul(gate_concat, g_d))
r_d = tf.nn.dropout(r_d, keep_prob)
r_q = tf.sigmoid(tf.matmul(gate_concat, g_q))
r_q = tf.nn.dropout(r_q, keep_prob)
combined_gated_glimpse = tf.concat([r_q * q_glimpse, r_d * d_glimpse], 1)
_, infer_state = infer_gru(combined_gated_glimpse, infer_state)
return tf.to_float(tf.sign(tf.abs(X))) * d_attention
开发者ID:veyvin,项目名称:tensorflow-learn,代码行数:52,代码来源:train.py
示例9: one_bp_iteration
def one_bp_iteration(self, xe_v2c_pre_iter, H_sumC_to_V, H_sumV_to_C, xe_0):
xe_tanh = tf.tanh(tf.to_double(tf.truediv(xe_v2c_pre_iter, [2.0])))
xe_tanh = tf.to_float(xe_tanh)
xe_tanh_temp = tf.sign(xe_tanh)
xe_sum_log_img = tf.matmul(H_sumC_to_V, tf.multiply(tf.truediv((1 - xe_tanh_temp), [2.0]), [3.1415926]))
xe_sum_log_real = tf.matmul(H_sumC_to_V, tf.log(1e-8 + tf.abs(xe_tanh)))
xe_sum_log_complex = tf.complex(xe_sum_log_real, xe_sum_log_img)
xe_product = tf.real(tf.exp(xe_sum_log_complex))
xe_product_temp = tf.multiply(tf.sign(xe_product), -2e-7)
xe_pd_modified = tf.add(xe_product, xe_product_temp)
xe_v_sumc = tf.multiply(self.atanh(xe_pd_modified), [2.0])
xe_c_sumv = tf.add(xe_0, tf.matmul(H_sumV_to_C, xe_v_sumc))
return xe_v_sumc, xe_c_sumv
开发者ID:liangfei-info,项目名称:Iterative-BP-CNN,代码行数:13,代码来源:BP_Decoder.py
示例10: _apply
def _apply(self, grad, var, indices=None):
lr = tf.cast(self._learning_rate_tensor, var.dtype.base_dtype)
m = self.get_slot(var, "m")
# m_t = beta1 * m + (1 - beta1) * g_t
beta1_t = tf.cast(self._beta1_t, var.dtype.base_dtype)
m_scaled_g_values = grad * (1 - beta1_t)
m_t = tf.assign(m, m * beta1_t, use_locking=self._use_locking)
with tf.control_dependencies([m_t]):
m_t = self._assign_add(m, updates=m_scaled_g_values, indices=indices)
# update = lr * grad * where(...)
m_gathered = self._gather(m_t, indices=indices)
ones = tf.ones_like(grad)
update = lr * grad * tf.where(tf.equal(tf.sign(m_gathered), tf.sign(grad)), ones, ones * self._decrease_factor)
var_update = self._assign_sub(ref=var, updates=update, indices=indices)
return tf.group(*[var_update, m_t])
开发者ID:rwth-i6,项目名称:returnn,代码行数:15,代码来源:TFUpdater.py
示例11: _apply_dense
def _apply_dense(self, grad, var):
lr_t = math_ops.cast(self._lr_t, var.dtype.base_dtype)
alpha_t = math_ops.cast(self._alpha_t, var.dtype.base_dtype)
beta_t = math_ops.cast(self._beta_t, var.dtype.base_dtype)
eps = 1e-7 # cap for moving average
m = self.get_slot(var, "m")
m_t = m.assign(tf.maximum(beta_t * m + eps, tf.abs(grad)))
var_update = state_ops.assign_sub(var, lr_t * grad * tf.exp(
tf.log(alpha_t) * tf.sign(grad) * tf.sign(m_t))) # Update 'ref' by subtracting 'value
# Create an op that groups multiple operations.
# When this op finishes, all ops in input have finished
return control_flow_ops.group(*[var_update, m_t])
开发者ID:jkhlot,项目名称:tensorflow-XNN,代码行数:15,代码来源:optimizer.py
示例12: build
def build(self):
""" tensorflow computation graph for transform """
graph = tf.Graph()
with graph.as_default():
self.inputs = tf.placeholder(tf.float32, shape=(None, self.max_atoms, 4))
atom_numbers = tf.cast(self.inputs[:, :, 0], tf.int32)
flags = tf.sign(atom_numbers)
flags = tf.cast(
tf.expand_dims(flags, 1) * tf.expand_dims(flags, 2), tf.float32)
coordinates = self.inputs[:, :, 1:]
if self.coordinates_in_bohr:
coordinates = coordinates * 0.52917721092
d = self.distance_matrix(coordinates, flags)
d_radial_cutoff = self.distance_cutoff(d, self.radial_cutoff, flags)
d_angular_cutoff = self.distance_cutoff(d, self.angular_cutoff, flags)
radial_sym = self.radial_symmetry(d_radial_cutoff, d, atom_numbers)
angular_sym = self.angular_symmetry(d_angular_cutoff, d, atom_numbers,
coordinates)
self.outputs = tf.concat(
[
tf.cast(tf.expand_dims(atom_numbers, 2), tf.float32), radial_sym,
angular_sym
],
axis=2)
return graph
开发者ID:ktaneishi,项目名称:deepchem,代码行数:25,代码来源:transformers.py
示例13: __graph__
def __graph__():
"""Building the inference graph"""
with tf.name_scope('input'):
# [BATCH_SIZE, NUM_FEATURES]
x_input = tf.placeholder(dtype=tf.float32, shape=[None, self.num_features], name='x_input')
# [BATCH_SIZE]
y_input = tf.placeholder(dtype=tf.uint8, shape=[None], name='y_input')
# [BATCH_SIZE, NUM_CLASSES]
y_onehot = tf.one_hot(indices=y_input, depth=self.num_classes, on_value=1, off_value=-1,
name='y_onehot')
learning_rate = tf.placeholder(dtype=tf.float32, name='learning_rate')
with tf.name_scope('training_ops'):
with tf.name_scope('weights'):
weight = tf.get_variable(name='weights',
initializer=tf.random_normal([self.num_features, self.num_classes],
stddev=0.01))
self.variable_summaries(weight)
with tf.name_scope('biases'):
bias = tf.get_variable(name='biases', initializer=tf.constant([0.1], shape=[self.num_classes]))
self.variable_summaries(bias)
with tf.name_scope('Wx_plus_b'):
output = tf.matmul(x_input, weight) + bias
tf.summary.histogram('pre-activations', output)
with tf.name_scope('svm'):
regularization = tf.reduce_mean(tf.square(weight))
hinge_loss = tf.reduce_mean(tf.square(tf.maximum(tf.zeros([self.batch_size, self.num_classes]),
1 - tf.cast(y_onehot, tf.float32) * output)))
with tf.name_scope('loss'):
loss = regularization + self.svm_c * hinge_loss
tf.summary.scalar('loss', loss)
optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate).minimize(loss)
with tf.name_scope('accuracy'):
predicted_class = tf.sign(output)
predicted_class = tf.identity(predicted_class, name='prediction')
with tf.name_scope('correct_prediction'):
correct = tf.equal(tf.argmax(predicted_class, 1), tf.argmax(y_onehot, 1))
with tf.name_scope('accuracy'):
accuracy = tf.reduce_mean(tf.cast(correct, 'float'))
tf.summary.scalar('accuracy', accuracy)
merged = tf.summary.merge_all()
self.x_input = x_input
self.y_input = y_input
self.y_onehot = y_onehot
self.learning_rate = learning_rate
self.loss = loss
self.optimizer = optimizer
self.output = output
self.predicted_class = predicted_class
self.accuracy = accuracy
self.merged = merged
开发者ID:TaihuLight,项目名称:wisconsin-breast-cancer,代码行数:60,代码来源:svm.py
示例14: triangle_wave
def triangle_wave(frequency):
"""Emit a triangle wave at the given frequency."""
xs = tf.reshape(tf.range(_samples(), dtype=tf.float32), [1, _samples(), 1])
ts = xs / FLAGS.sample_rate
#
# A triangle wave looks like this:
#
# /\ /\
# / \ / \
# \ / \ /
# \/ \/
#
# If we look at just half a period (the first four slashes in the
# diagram above), we can see that it looks like a transformed absolute
# value function.
#
# Let's start by computing the times relative to the start of each
# half-wave pulse (each individual "mountain" or "valley", of which
# there are four in the above diagram).
half_pulse_index = ts * (frequency * 2)
half_pulse_angle = half_pulse_index % 1.0 # in [0, 1]
#
# Now, we can see that each positive half-pulse ("mountain") has
# amplitude given by A(z) = 0.5 - abs(z - 0.5), and then normalized:
absolute_amplitude = (0.5 - tf.abs(half_pulse_angle - 0.5)) / 0.5
#
# But every other half-pulse is negative, so we should invert these.
half_pulse_parity = tf.sign(1 - (half_pulse_index % 2.0))
amplitude = half_pulse_parity * absolute_amplitude
#
# This is precisely the desired result, so we're done!
return amplitude
开发者ID:jlewi,项目名称:tensorboard,代码行数:32,代码来源:audio_demo.py
示例15: binomial_sampling
def binomial_sampling(self, pr):
"""
Binomial sampling of hidden units activations using a rejection method.
Basic mechanics:
1) Extract a random number from a uniform distribution (g) and compare it with
the unit's probability (pr)
2) Choose 0 if pr<g, 1 otherwise. It is convenient to implement this condtion using
the relu function.
Args:
pr (tensor, float32): input conditional probability
g (np.array, float32): uniform probability used for comparison
Returns:
h_sampled (tensor, float32): sampled units. The value is 1 if pr>g and 0 otherwise.
"""
np.random.seed(self.seed)
# sample from a Bernoulli distribution with same dimensions as input distribution
g = tf.convert_to_tensor(np.random.uniform(size=pr.shape[1]), dtype=tf.float32)
# sample the value of the hidden units
h_sampled = tf.nn.relu(tf.sign(pr - g))
return h_sampled
开发者ID:David-Li-L,项目名称:recommenders,代码行数:29,代码来源:rbm.py
示例16: retrieve_seq_length_op
def retrieve_seq_length_op(data):
"""An op to compute the length of a sequence. 0 are masked. """
with tf.name_scope('GetLength'):
used = tf.sign(x=tf.reduce_max(tf.abs(data), axis=2))
length = tf.reduce_sum(input_tensor=used, axis=1)
length = tf.cast(x=length, dtype=tf.int32)
return length
开发者ID:AlexMikhalev,项目名称:polyaxon,代码行数:7,代码来源:recurrent.py
示例17: loss
def loss(logits, labels):
"""Calculates Mean Pixel Error.
Args:
logits: Logits from inference().
labels: Labels from distorted_inputs or inputs(). 1-D tensor
of shape [batch_size]
Returns:
Loss tensor of type float.
"""
labelValidity = tf.sign(labels, name='label_validity')
minop = tf.sub(logits, labels, name='Diff_Op')
absop = tf.abs(minop, name='Abs_Op')
lossValues = tf.mul(labelValidity, absop, name='lossValues')
loss_mean = tf.reduce_mean(lossValues, name='MeanPixelError')
tf.add_to_collection('losses', loss_mean)
return tf.add_n(tf.get_collection('losses'), name='total_loss'), loss_mean
开发者ID:bdutta19,项目名称:deeppose,代码行数:25,代码来源:LSPModels.py
示例18: encode
def encode(self, x, noise):
x = tf.to_float(x)
# we can't use tf.pow(..., 8.0) because of a high-error approximation
# on TPU. Instead we square three times.
x = tf.sign(x) * tf.square(tf.square(tf.square(tf.abs(x) * 128.0)))
x = _to_bfloat16_unbiased(x, noise)
return x
开发者ID:kltony,项目名称:tensor2tensor,代码行数:7,代码来源:quantization.py
示例19: _non_linear_grad
def _non_linear_grad(cls, op, grad):
LRP.logger.debug("Computing non-linear gradient with activation type {}".format(op.type))
op_out = op.outputs[0]
op_in = op.inputs[0]
stabilizer_epsilon = cls._eps * tf.sign(op_in)
op_in += stabilizer_epsilon
return grad * op_out / op_in
开发者ID:ashishyadavppe,项目名称:Skater,代码行数:7,代码来源:relevance_scorer.py
示例20: __call__
def __call__(self,x,keep_prob=1.0,seq_length=None): #__call__ is very efficient when the state of instance changes frequently
with tf.variable_scope(self.name,reuse = self.reuse) as vs:
self.fw_cell =tf.contrib.rnn.LSTMCell(self.cell_size,state_is_tuple=True,reuse=tf.get_variable_scope().reuse)
self.fw_cell1 =tf.contrib.rnn.LSTMCell(self.cell_size,state_is_tuple=True,reuse=tf.get_variable_scope().reuse)
self.bw_cell =tf.contrib.rnn.LSTMCell(self.cell_size,state_is_tuple=True,reuse=tf.get_variable_scope().reuse)
self.bw_cell1 =tf.contrib.rnn.LSTMCell(self.cell_size,state_is_tuple=True,reuse=tf.get_variable_scope().reuse)
self.fw_cells = tf.contrib.rnn.MultiRNNCell([self.fw_cell,self.fw_cell1],state_is_tuple=True)
self.bw_cells = tf.contrib.rnn.MultiRNNCell([self.bw_cell,self.bw_cell1],state_is_tuple=True)
if seq_length ==None: #get the real sequence length (suppose that the padding are zeros)
used = tf.sign(tf.reduce_max(tf.abs(x),reduction_indices=2))
seq_length = tf.cast(tf.reduce_sum(used,reduction_indices=1),tf.int32)
lstm_out,_,_ = tf.contrib.rnn.static_bidirectional_rnn(self.fw_cells,self.bw_cells,tf.unstack(tf.transpose(x,[1,0,2])),dtype=tf.float32,sequence_length=seq_length)
lstm_out = tf.transpose(tf.stack(lstm_out),[1,0,2])
print 'lstm_out: ',lstm_out
#shape(lstm_out) = (self.batch_size,sequence_length,2*cell_size)
#if keep_prob < 1.:
# lstm_out = tf.nn.dropout(lstm_out,keep_prob)
if self.reuse is None:
self.trainable_weights = vs.global_variables()
self.reuse =True
return lstm_out,seq_length
开发者ID:wujsAct,项目名称:TeachingMachineReadAndComprehend,代码行数:31,代码来源:layers.py
注:本文中的tensorflow.sign函数示例由纯净天空整理自Github/MSDocs等源码及文档管理平台,相关代码片段筛选自各路编程大神贡献的开源项目,源码版权归原作者所有,传播和使用请参考对应项目的License;未经允许,请勿转载。 |
请发表评论