1、ReLU 函数
2、sigmod函数
a = tf.linspace(-10., 10., 10)
with tf.GradientTape() as tape:
tape.watch(a) # 如果将a定义为Variable,此句话可以省略
y = tf.sigmoid(a)
grads = tape.gradient(y, [a])
3、tanh函数
4、leakyReLU函数
1、均方差损失函数
with tf.GradientTape() as tape:
tape.watch([w, b])
logits = tf.sigmoid(x@w+b)
loss = tf.reduce_mean(tf.losses.MSE(y, logits))
grads = tape.gradient(loss, [w, b])
print('w grad:', grads[0])
print('b grad:', grads[1])
2 、交叉熵函数梯度
with tf.GradientTape() as tape:
tape.watch([w, b])
logits = (x@w+b)
loss = tf.reduce_mean(tf.losses.categorical_crossentropy(y, logits, rom_logits=True))
#rom_logits=True 可以省略softmax那一步
grads = tape.gradient(loss, [w, b])
3、反向传播推倒:略
4、函数优化
def himmelblau(x):
# himmelblau 函数实现
return (x[0] ** 2 + x[1] - 11) ** 2 + (x[0] + x[1] ** 2 - 7) ** 2
x = tf.constant([-4.,0.])
for step in range(200):
with tf.GradientTape() as tape:
tape.watch(x)
y = himmelblau(x)
grads = tape.gradient(y,x)
x -= 0.01*grads
if step % 20 == 0:
print('step{}:x = {},f(x) = {}'.format(step, x, y.numpy()))
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import datasets, layers, optimizers, Sequential, metrics
import os
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
def preprocess(x, y):
"""
:param x:
:param y:
:return:
"""
# [b, 28, 28], [b]
x = tf.cast(x, dtype=tf.float32) / 255.
y = tf.cast(y, dtype=tf.int32)
y = tf.one_hot(y, depth=10)
return x, y
def main():
(x, y), (x_test, y_test) = datasets.fashion_mnist.load_data()
db = tf.data.Dataset.from_tensor_slices((x, y))
db = db.map(preprocess).shuffle(10000).batch(128)
db_test = tf.data.Dataset.from_tensor_slices((x_test, y_test))
db_test = db_test.map(preprocess).batch(128)
#db_iter = iter(db)
#sample = next(db_iter)
# print(sample[1].shape)
model = Sequential([
layers.Dense(256, activation=tf.nn.relu),
layers.Dense(128, activation=tf.nn.relu),
layers.Dense(64, activation=tf.nn.relu),
layers.Dense(32, activation=tf.nn.relu),
layers.Dense(10, activation=tf.nn.relu),
])
#model.build(input_shape=[None, 28 * 28])
#model.summary()
optimizer = optimizers.Adam(lr=1e-3)
for epock in range(30):
for step, (x, y) in enumerate(db):
x = tf.reshape(x, [-1, 28 * 28])
with tf.GradientTape() as tape:
logits = model(x)
loss_mse = tf.reduce_mean(tf.losses.MSE(y, logits))
loss_cro = tf.reduce_mean(tf.losses.categorical_crossentropy(y, logits, from_logits=True))
grads = tape.gradient(loss_mse, model.trainable_variables)
optimizer.apply_gradients(zip(grads, model.trainable_variables))
if step % 200 == 0:
print(epock, step, 'loss', float(loss_mse), float(loss_cro))
total_num = 0
total_correct = 0
for x, y in db_test:
x = tf.reshape(x, [-1, 28 * 28])
logit = model(x)
prob = tf.nn.softmax(logit, axis=1)
pred = tf.argmax(prob, axis=1)
pred = tf.cast(pred, tf.int32)
y = tf.argmax(y, axis=1)
y = tf.cast(y, dtype=tf.int32)
correct = tf.equal(pred, y)
correct = tf.reduce_sum(tf.cast(correct, dtype=tf.int32))
total_correct += int(correct)
total_num += x.shape[0]
acc = total_correct / total_num
print(acc)
if __name__ == '__main__':
main()
命令:tensorboard --logdir log
def plot_to_image(figure):
"""Converts the matplotlib plot specified by 'figure' to a PNG image and
returns it. The supplied figure is closed and inaccessible after this call."""
# Save the plot to a PNG in memory.
buf = io.BytesIO()
plt.savefig(buf, format='png')
# Closing the figure prevents it from being displayed directly inside
# the notebook.
plt.close(figure)
buf.seek(0)
# Convert PNG buffer to TF image
image = tf.image.decode_png(buf.getvalue(), channels=4)
# Add the batch dimension
image = tf.expand_dims(image, 0)
return image
def image_grid(images):
"""Return a 5x5 grid of the MNIST images as a matplotlib figure."""
# Create a figure to contain the plot.
figure = plt.figure(figsize=(10, 10))
for i in range(16):
# Start next subplot.
plt.subplot(4, 4, i + 1, title='name')
plt.xticks([])
plt.yticks([])
plt.grid(False)
plt.imshow(images[i])
return figure
with summary_writer.as_default():
tf.summary.scalar('test-acc', float(total_correct / total_num), step=step)
tf.summary.image("val-onebyone-images:", val_images, max_outputs=25, step=step)
val_images = tf.reshape(val_images, [-1, 28, 28])
figure = image_grid(val_images)
tf.summary.image('val-images:', plot_to_image(figure), step=step)
tf.summary.scalar('loss', float(loss_mse), step=epock)
tf.summary.scalar('acc', float(total_correct / total_num), step=epock)