tensorflow 2.0学习教程一

最新推荐文章于 2023-09-07 12:49:46 发布

地主家的小兵将

最新推荐文章于 2023-09-07 12:49:46 发布

阅读量438

点赞数

CC 4.0 BY-SA版权

分类专栏： python 文章标签： tensorflow

本文链接：https://blog.youkuaiyun.com/bingheshidai_1234/article/details/94619660

python 专栏收录该内容

38 篇文章

订阅专栏

这篇博客介绍了如何使用TensorFlow 2.0进行深度学习，特别关注了数据集的获取和使用。提供的链接指向了一个包含详细教程的数据集。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

import matplotlib as mpl
import matplotlib.pyplot as plt
import numpy as np
import sklearn
import pandas as pd
import os
import sys
import time
import tensorflow as tf
from tensorflow import keras

'''
sys.version_info(major=3, minor=5, micro=4, releaselevel='final', serial=0)
matplotlib 3.0.3
numpy 1.16.2
pandas 0.24.2
sklearn 0.20.3
tensorflow 1.11.0
tensorflow.keras 2.1.6-tf

'''
print(tf.__version__)
print(sys.version_info)
for module in mpl, np, pd, sklearn, tf, keras:
    print(module.__name__, module.__version__)
def load_data(path,files):
    import gzip
    import numpy as np
    paths=[path+each for each in files]
    with gzip.open(paths[0],'rb') as lbpath:
        train_labels = np.frombuffer(lbpath.read(),np.uint8,offset=8)
    with gzip.open(paths[1],'rb') as imgpath:
        train_images = np.frombuffer(imgpath.read(),np.uint8,offset=16).reshape(len(train_labels),28,28)
    with gzip.open(paths[2],'rb') as lbgpath:
        test_labels = np.frombuffer(lbgpath.read(),np.uint8,offset=8)
    with gzip.open(paths[3],'rb') as imgpath:
        test_images = np.frombuffer(imgpath.read(),np.uint8,offset=16).reshape(len(test_labels),28,28)
    return (train_images,train_labels),(test_images,test_labels)
path='fashion-mnist/'
files=['train-labels-idx1-ubyte.gz','train-images-idx3-ubyte.gz','t10k-labels-idx1-ubyte.gz','t10k-images-idx3-ubyte.gz']

(x_train_all, y_train_all), (x_test, y_test) = load_data(path,files)
x_valid, x_train = x_train_all[:5000], x_train_all[5000:]
y_valid, y_train = y_train_all[:5000], y_train_all[5000:]
#做数据归一化处理  使其满足正太分布
# x = (x - u) / std
from  sklearn.preprocessing import StandardScaler
scaler = StandardScaler()
x_train_scaled =scaler.fit_transform(x_train.astype(np.float32).reshape(-1,1)).reshape(-1,28,28)
x_valid_scaled =scaler.transform(x_valid.astype(np.float32).reshape(-1,1)).reshape(-1,28,28)
x_test_scaled =scaler.transform(x_test.astype(np.float32).reshape(-1,1)).reshape(-1,28,28)
# model=keras.models.Sequential([
#     keras.layers.Flatten(input_shape=[28,28]), #展平 28*28
#     keras.layers.Dense(300,activation='relu'),
#     keras.layers.Dense(100,activation='relu'),
#     keras.layers.Dense(10,activation='softmax')
# ])
model=keras.models.Sequential()
model.add(keras.layers.Flatten(input_shape=[28,28]))
for _ in range(20):
    model.add(keras.layers.Dense(100,activation='relu'))
    # model.add(keras.layers.BatchNormalization())  #（1）加速收敛（2）控制过拟合，可以少用或不用Dropout和正则（3）降低网络对初始化权重不敏感（4）允许使用较大的学习率
# model.add(keras.layers.AlphaDropout(rate=0.5)) #减少过拟合
model.add(keras.layers.Dense(10,activation='softmax'))

#目标函数
model.compile(loss="sparse_categorical_crossentropy",optimizer = "sgd",metrics = ["accuracy"])
logdir='./dnn-bn-callbacks'
if not os.path.exists(logdir):
    os.mkdir(logdir)
output_model_file = os.path.join(logdir,'fashion_mnist_model.h5')
callbacks=[
    keras.callbacks.TensorBoard(logdir),
    keras.callbacks.ModelCheckpoint(output_model_file,save_best_only=True),
    keras.callbacks.EarlyStopping(patience=5,min_delta=1e-3),
]
history=model.fit(x_train_scaled,y_train,epochs=20,validation_data=(x_valid_scaled,y_valid),callbacks=callbacks)
def plot_learn_curves(history):
    pd.DataFrame(history.history).plot(figsize=(8,5))
    plt.grid(True)
    plt.ylim(0,1)
    plt.show()
plot_learn_curves(history)
model.evaluate(x_test_scaled,y_test)

数据集：链接：https://pan.baidu.com/s/1mI-JZW8042xikQ6NRlebUg
提取码：t8va