def test_lookup_positives(self):
scores = constant([[1., 0., 2.], [0., 1., 1.]])
clicks = constant([[2], [1]])
expected = [2., 1.]
positive_scores = lookup_positives(scores, clicks)
with self.test_session():
self.assertListEqual(positive_scores.eval().tolist(), expected)
def test_predict_scores(self):
features = {
'anchor_label':
constant([[0., 1], [1., 0.]]),
'label':
constant([[[0., 1.], [0., 1.], [0., 1.]],
[[1., 0.], [1., 0.], [1., 0.]]])
}
p = predict_scores(features)
with self.test_session() as sess:
sess.run(global_variables_initializer())
print(p.eval())
def test_create_diffs(self):
scores = constant([[1., 0., 2.], [0., 1., 1.]])
positive_scores = [2., 1.]
expected = [[1., 2., 0.], [1., 0., 0.]]
diffs = create_diffs(positive_scores, scores)
with self.test_session():
self.assertListEqual(diffs.eval().tolist(), expected)
def _calculate_supervised_similarities(y_true) -> Tensor:
"""
Calculates the target supervised similarities.
Performs a tensorflow nested loop, in order to compare the values of y_true for range(batch_size).
:param y_true: the y_true value.
:return: Tensor containing the target supervised similarities.
"""
# Get the batch size.
batch_size = shape(y_true)[0]
# Initialize outer loop index.
i = constant(0)
# Initialize symmetric supervised similarity matrix targets.
target_similarity = zeros((batch_size, batch_size))
def outer_loop_condition(_i, _batch_size, _y_true, _target_similarity):
"""Define outer loop condition."""
return less(_i, _batch_size)
def outer_loop_body(_i, _batch_size, _y_true, _target_similarity):
"""Define outer loop body."""
# Initialize inner loop index.
j = constant(0)
def inner_loop_condition(_i, _j, _y_true, _target_similarity):
"""Define inner loop condition."""
return less(_j, _batch_size)
def inner_loop_body(_i, _j, _y_true, _target_similarity):
"""Define inner loop body."""
if _y_true[_i] == _y_true[_j]:
_target_similarity[_i, _j] = 1
return _i, _j + 1, _y_true, _target_similarity
# Begin inner while loop.
_, j, _, _target_similarity = while_loop(
inner_loop_condition, inner_loop_body,
[_i, j, _y_true, _target_similarity])
return _i + 1, _batch_size, _y_true, _target_similarity
# Begin outer while loop.
i, _, _, target_similarity = while_loop(
outer_loop_condition, outer_loop_body,
[i, batch_size, y_true, target_similarity])
return target_similarity
def outer_loop_body(_i, _batch_size, _y_true, _target_similarity):
"""Define outer loop body."""
# Initialize inner loop index.
j = constant(0)
def inner_loop_condition(_i, _j, _y_true, _target_similarity):
"""Define inner loop condition."""
return less(_j, _batch_size)
def inner_loop_body(_i, _j, _y_true, _target_similarity):
"""Define inner loop body."""
if _y_true[_i] == _y_true[_j]:
_target_similarity[_i, _j] = 1
return _i, _j + 1, _y_true, _target_similarity
# Begin inner while loop.
_, j, _, _target_similarity = while_loop(
inner_loop_condition, inner_loop_body,
[_i, j, _y_true, _target_similarity])
return _i + 1, _batch_size, _y_true, _target_similarity
def run():
# 创建一个变量, 初始化为标量 0.
state = Variable(0, name="counter")
# 创建一个 op, 其作用是使 state 增加 1
one = constant(1)
new_value = add(state, one)
update = assign(state, new_value)
# 启动图后, 变量必须先经过`初始化` (init) op 初始化,
# 首先必须增加一个`初始化` op 到图中.
init_op = initialize_all_variables()
# 启动图, 运行 op
with Session() as sess:
# 运行 'init' op
sess.run(init_op)
# 打印 'state' 的初始值
print(sess.run(state))
# 运行 op, 更新 'state', 并打印 'state'
for _ in range(3):
sess.run(update)
print(sess.run(state))
from BBDATA import *
import tensorflow.python as tf
from cnn_utils import save_model
import matplotlib.pyplot as plt
train_times = 50000
base_path = "/Users/coorchice/Desktop/ML/model/ml/BreadBasket/"
save_path = base_path + str(train_times) + "/"
BBDATA = read_datas('data/')
x_data = tf.placeholder(tf.float32, [None, 135])
y_data = tf.placeholder(tf.float32, [None])
W = tf.Variable(tf.truncated_normal([135, 1], stddev=0.1))
b = tf.Variable(tf.constant(0.1, shape=[1]))
y = tf.nn.relu(tf.matmul(x_data, W) + b)
# 按照交叉熵公式计算交叉熵
with tf.name_scope('loss'):
# cross_entropy = -tf.reduce_sum(y_data * tf.log(y))
cross_entropy = tf.reduce_mean((tf.square((y - y_data))))
tf.scalar_summary('loss', cross_entropy)
# init_lr = 0.00001
lr = tf.Variable(0.00005, trainable=False)
# global_step = tf.Variable(0., trainable=False)
# lr = tf.train.exponential_decay(init_lr, global_step=global_step, decay_steps=10000, decay_rate=0.5, staircase=True)
# 使用梯度下降法不断的调整变量，寻求最小的交叉熵
# 此处使用梯度下降法以0.01的学习速率最小化交叉熵
def bias_variable(shape):
initial = tf.constant(0.1, shape=shape)
return tf.Variable(initial)
# -*- coding:utf-8 -*- ''' @Author: zzx @E-mail: [email protected] @File: 对多元函数求导.py @CreateTime: 2020/7/21 15:13 ''' import tensorflow.python as tf # 2、多元函数求导 X = tf.constant([[1., 2.], [3., 4.]]) y = tf.constant([[1.], [2.]]) # 函数参数,初始化参数随便定义 w = tf.Variable(initial_value=[[1.], [2.]]) b = tf.Variable(initial_value=1.) # 在这里可以执行自动求导 with tf.GradientTape() as tape: L = 0.5 * tf.reduce_sum(tf.square(tf.matmul(X, w) + b - y)) w_grad, b_grad = tape.gradient(L, [w, b]) print("".format(L.numpy(), w_grad.numpy(), b_grad.numpy()))
def test_consant():
t = tf.ones((16, 10))
b = tf.constant(tf.zeros((t.shape[0], t.shape[1])),
name='b') # b [in_caps,out_caps]