TensorFlow1:
%matplotlib inline
import matplotlib.pyplot as plt
import numpy as np
import tensorflow as tf
np.random.seed(5)
x_data = np.linspace(-1,1,100)
#help(np.random.randn)
#print(np.random.randn(*x_data.shape))
#np.random.randn(*x_data.shape) *0.4
#y_data = 2*x_data+1+np.random.randn(*x_data.shape) *0.4
#plt.scatter(x_data,y_data)
#plt.plot(x_data,2*x_data+1,color='red',linewidth=3)
x = tf.compat.v1.placeholder(tf.float32,name='x')
y = tf.compat.v1.placeholder(tf.float32,name='y')
def model(x,w,b):
return tf.multiply(x,w)+b
w = tf.Variable(1.0,name='w0')
b = tf.Variable(0.0,name='b0')
pred = model(x,w,b)
train_epochs = 10
learning_rate = 0.05
loss_function = tf.reduce_mean(tf.square(y-pred))
optimizer = tf.compat.v1.train.GradientDescentOptimizer(learning_rate).minimize(loss_function)
sess = tf.compat.v1.Session()
init = tf.compat.v1.global_variables_initializer()
sess.run(init)
for epoch in range(3):
for xs,ys in zip(x_data,y_data):
_,loss = sess.run([optimizer,loss_function],feed_dict={x:xs,y:ys})
b0temp = b.eval(session=sess)
w0temp = w.eval(session=sess)
plt.plot(x_data,w0temp*x_data+b0temp)
print("w:",sess.run(w))
print("b:",sess.run(b))
TensorFlow2:
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
print(tf.__version__)
x_data = np.linspace(-1,1,100)
y_data = 2 * x_data + 1 +np.random.randn(*x_data.shape) * 0.4
w = tf.Variable(1.0,tf.float32)
b = tf.Variable(0.0,tf.float32)
def model(x,w,b):
return tf.multiply(w,x) + b
def loss(x,w,b,y):
pred = model(x,w,b)
square = tf.square(pred-y)
return tf.reduce_mean(square)
def grad(x,w,b,y):
with tf.GradientTape() as tape:
loss_ = loss(x,w,b,y)
return tape.gradient(loss_,[w,b])
step = 0
loss_list = []
disply_step = 10
training_epochs = 10
learning_rate = 0.05
#BGD
for epochs in range(training_epochs):
loss_ = loss(x_data,w,b,y_data)
loss_list.append(loss_)
print(loss_.numpy())
delta_w,delta_b = grad(xs,w,b,ys)
change_w = delta_w * learning_rate
change_b = delta_b * learning_rate
w.assign_sub(change_w)
b.assign_sub(change_b)
step+=1
if step % disply_step == 0:
print(step,"--->",loss_)
#SGD
#for epochs in range(training_epochs):
#for xs,ys in zip(x_data,y_data):
#loss_ = loss(xs,w,b,ys)
#loss_list.append(loss_)
#delta_w,delta_b = grad(xs,w,b,ys)
#change_w = delta_w * learning_rate
#change_b = delta_b * learning_rate
#w.assign_sub(change_w)
#b.assign_sub(change_b)
#step+=1
#if step % disply_step == 0:
#print(step,"--->",loss_)
print("w",w.numpy())
print("b",b.numpy())
#plt.plot(x_data,x_data*w.numpy() + b.numpy())
plt.plot(loss_list)