3/25

- np.random.RandomState() if state is None else state

# 방법1
if state is None:
  np.random.RandomState()
else:
  state

# 방법2
np.random.RandomState() if state is None else state

 

- .real

a = 1+2j
a.real # 1.0
a.imag # 2.0

- 이중 def

def abc(x):
    def ab(*arg, **kwargs)
        return x(*arg, **kwargs)
    return ab

 

 

- 이중 for문 + 람다함수

# 방법1
flatten = lambda l: [item for sublist in l for item in sublist]

# 방법2
for sublist in l :
  for item in sublist :
    item

 

- assert == if

assert sample_rate==16000

 

- np.row_stack() == np.vstack()

 

- np.expand_dims(x, axis = 2)

a = np.array([1,2])
a.shape # (2,)

a.expand_dims(a, axis=0)
a.shape # (1,2)

 

- tf의 Variance Scope 경로메커니즘
= tf.get_variance(name, shape, initializer) + tf.variance_scope(scope_name)

initializer =
tf.constant_initializer(value)
+ tf.random_uniform_initializer(a, b) # [a, b]를 균일하게 초기화 합니다,
+ tf.random_normal_initializer(mean, stddev) # 주어진 평균 및 표준 편차로 정규 분포에서 초기화합니다.

 

ex)
def image_filter():
   with tf.variance_scope('a'):
            tf.variance_scope('b'):
               tf.get_variance('weight') # 'a/b/weight'

with tf.variable_scope("image_filters") as scope:
    result1 = my_image_filter(image1)
    scope.reuse_variables()
    result2 = my_image_filter(image2)

 

 

def my_image_filter(input_images):
    with tf.variable_scope("conv1"):
        # Variables created here will be named "conv1/weights", "conv1/biases".
        relu1 = conv_relu(input_images, [5, 5, 32, 32], [32])
    with tf.variable_scope("conv2"):
        # Variables created here will be named "conv2/weights", "conv2/biases".
        return conv_relu(relu1, [5, 5, 32, 32], [32])

# 연속호출 불가
result1 = my_image_filter(image1)
result2 = my_image_filter(image2)
# Raises ValueError(... conv1/weights already exists ...)

# 연속호출가능 - .reuse_variables()
with tf.variable_scope("image_filters") as scope:
    result1 = my_image_filter(image1)
    scope.reuse_variables()
    result2 = my_image_filter(image2)

# op 함수 name
with tf.variable_scope("foo"):
    x = 1.0 + tf.get_variable("v", [1])
assert x.op.name == "foo/add"

 

 

 

 

 

읽은것 :

https://tensorflowkorea.gitbooks.io/tensorflow-kr/content/g3doc/how_tos/variables/

변수 · 텐서플로우 문서 한글 번역본

https://tensorflowkorea.gitbooks.io/tensorflow-kr/content/g3doc/how_tos/variable_scope/

공유 변수 · 텐서플로우 문서 한글 번역본