TensorFlow Code Snippets

tensor creating and modifying placeholder

Https://www.tensorflow.org/api_docs/python/tf/placeholder

Tf.placeholder (
    dtype,
    shape=none,
    name=none
)

element Access Access specific elements by mask

Import NumPy as NP
import TensorFlow as TF;

A = Tf.placeholder (Dtype = tf.float32)
mask = a > 1;
Mask.set_shape ([None, none]); # The shape of the mask must not being none.
b = Tf.placeholder (Dtype = tf.float32)
c = Tf.boolean_mask (b, mask);
With TF. Session () as sess:
    a_val = Np.ones ((3,))
    a_val[1] = 3;
    B_val = a_val
    c_val = Sess.run (c, feed_dict = {a:a_val, b:b_val})
    print C_val

Shape Ascending Dimension

input = tf.expand_dims (
    input,
    axis=none,
    name=none
)

Set Shape

A common scenario is that after an operation, the shape information of the tensor becomes none, but the operator (or we) knows which dim shape information is certain, such as the image shape below, and the third dimension can determine that it will remain unchanged at 3. This information is set up here, you can save a lot of trouble for the later programming.

# Crop The image to the specified bounding box.
Cropped_image = Tf.slice (image, Bbox_begin, bbox_size)
# Restore The shape since the dynamic slice loses 3rd dimension .
Cropped_image.set_shape ([None, none, 3])

related to variable names tf.identity

Import TensorFlow as tf
a = tf. Variable (1, name = ' a ')
B = tf. Variable (2, name = ' B ') with
tf.name_scope (' SS '):
    C = a + b
    c = tf.identity (c, ' C ')
    #c = Tf.add (A, B, Nam e = ' c ')

print C.name

ss/c:0 Type Conversions

Fnmask = Tf.cast (Nmask, Dtype)

Basic Operations sum two variables added

Import TensorFlow as tf

a = tf. Variable (1, name = ' a ')
B = tf. Variable (2, name = ' B ')

C = Tf.add (A, b, name = ' C ')
C = A + b

all elements Plus and

scalar = tf.reduce_sum (tensor)

Max/min Maximum/small element in a tensor

Import NumPy as NP
import TensorFlow as TF;

A = Tf.placeholder (Dtype = tf.float32)
B = Tf.reduce_max (a) # Reduce_min with

TF. Session () as sess:
    a_val = Np.ones ((3, 3)
    a_val[0, 2] = B_val
    = Sess.run (b, feed_dict = {a:a_val}) 
  print B_val

100.0 two x tensor by element Max/small

Import NumPy as NP
import TensorFlow as TF;

A = Tf.placeholder (Dtype = tf.float32)
B = tf.placeholder (Dtype = tf.float32)
c = Tf.maximum (A, B) # minimum
WI TH TF.  Session () as sess:
    a_val = Np.ones ((3, 3)
    a_val[0, 2] = b_val
    = a_val
    c_val = Sess.run (c, feed_dict = {a:a_val, b:b_val})
    Print C_val

[1.1. 100.]
[1.1. 1.]
[1.1. 1.]] Bool and

Mask = Tf.logical_and (Mask1, Mask2);

Loss Weighted Loss

Loss_pos = Tf.losses.compute_weighted_loss (
    losses,
    weights=1.0,
    scope=none,
    loss_collection =TF. Graphkeys.losses
)

The average is calculated using a non-0 number in the weights. optimization Gradient Clip

Https://stackoverflow.com/questions/36498127/how-to-effectively-apply-gradient-clipping-in-tensor-flow

Optimizer = Tf.train.AdamOptimizer (learning_rate=learning_rate)
GVs = optimizer.compute_gradients (cost)
Capped_gvs = [(Tf.clip_by_value (Grad, -1., 1.), Var) for Grad, var in gvs]
train_op = optimizer.apply_gradients (capped _gvs)

But this clip cannot handle Nan:

Import NumPy as NP
import TensorFlow as tf
val = float (' nan ')
a = Np.ones ((3, 3)) * 0.5
a[0,0] = 2
a[0 , 1] = val
tf_a = tf. Variable (a)
Tf_b = Tf.clip_by_value (A,-1, 1) with
tf.device ('/cpu:0 '): With
    TF. Session () as Sess:
        Sess.run (Tf.global_variables_initializer ())
        print Sess.run ([tf_a, Tf_b])

[Array ([[2],  nan,  0.5],
       [0.5,  0.5,  0.5],
       [0.5,  0.5,  0.5]]), 
       array ([1.,  nan,  0.5],
       [0.5,  0.5,  0.5],
       [0.5,  0.5,  0.5]])

Use Tf.is_finite to process both Nan and infty

Import NumPy as NP
import TensorFlow as tf
val = float (' nan ')
a = Np.ones ((3, 3)) * 0.5
a[0,0] = np.infty< c4/>a[0,1] = val
tf_a = tf. Variable (a)
#tf_b = Tf.clip_by_value (A,-1, 1)
mask = tf.is_finite (tf_a)
tf_c = Tf.where (Mask, tf_a, Tf.zeros_like (a)) with
tf.device ('/cpu:0 '): With
    TF. Session () as Sess:
        Sess.run (Tf.global_variables_initializer ())
        print Sess.run ([tf_a, Tf_c])

[Array ([INF,  nan,  0.5],
       [0.5,  0.5,  0.5],
       [0.5,  0.5,  0.5]]), 
 Array ([[0.,  0.,  0.5],
       [0.5,  0.5,  0.5],
       [0.5,  0.5,  0.5]])

Py_func

Https://www.tensorflow.org/api_docs/python/tf/py_func

Import NumPy as NP
import TensorFlow as TF

def add (A, B):
    return a + B;


Def Tf_add (A, B):
    return Tf.py_func (Add, [A, b], tf.float32);


Ta = tf.placeholder (Dtype = tf.float32)
TB = tf.placeholder (Dtype = tf.float32)
TC = Tf_add (TA, TB);
With TF. Session () as sess:
    a_val = Np.ones ((2, 2))
    B_val = A_val * 2;
    Print Sess.run ([TC], feed_dict = {ta:a_val, tb:b_val})

Py_func receives three parameters: The function object, the TF type to return the value of the tensor, to be passed in. If the type parameter is a list, for example [Tf.float32], the return value is also a list.