>- 本文为[🔗365天深度学习训练营](https://mp.weixin.qq.com/s/k-vYaC8l7uxX51WoypLkTw) 中的学习记录博客
>- 参考文章地址： [🔗深度学习100例-卷积神经网络（CNN）天气识别 | 第5天](https://mtyjkh.blog.csdn.net/article/details/117186183)
配置环境
1.先用conda 创建一个新环境
conda create -n sdxx python=3.6.5
conda activate sdxx
2.安装所需要的软件
conda install -c conda-forge jupyter-offlinenotebook
conda install -c conda-forge tensorflow##没装上，换PYPI试一试
pip install tensorflow
pip install matplotlib
pip3 install jupyter
pip list
然后打开jupyter notebook
jupyter notebook
页面如下：
点击new，创建一个python3的新的页面
3.导入数据
import matplotlib.pyplot as plt
import os,PIL
# 设置随机种子尽可能使结果可以重现
import numpy as np
np.random.seed(1)
# 设置随机种子尽可能使结果可以重现
import tensorflow as tf
tf.random.set_seed(1)
from tensorflow import keras
from tensorflow.keras import layers,models
import pathlib
运行情况大概是这样
data_dir = "/scratch/2022-08-06/med-zhouh/sdxx/5day/weather_photos/"
data_dir = pathlib.Path(data_dir)
3.查看数据
数据集一共分为cloudy、rain、shine、sunrise四类，分别存放于weather_photos文件夹中以各自名字命名的子文件夹中。
image_count = len(list(data_dir.glob('*/*.jpg')))
print("",image_count)
roses = list(data_dir.glob('sunrise/*.jpg'))
PIL.Image.open(str(roses[0]))
输出结果是这样的
继续第二部分的学习
1.加载数据
使用image_dataset_from_directory方法将磁盘中的数据加载到tf.data.Dataset中
batch_size = 32
img_height = 180
img_width = 180
train_ds = tf.keras.preprocessing.image_dataset_from_directory(
data_dir,
validation_split=0.2,
subset="training",
seed=123,
image_size=(img_height, img_width),
batch_size=batch_size)
“”"
val_ds = tf.keras.preprocessing.image_dataset_from_directory(
data_dir,
validation_split=0.2,
subset="validation",
seed=123,
image_size=(img_height, img_width),
batch_size=batch_size)
得到结果如下：
Found 1125 files belonging to 4 classes.
Using 225 files for validation.
我们可以通过class_names输出数据集的标签。标签将按字母顺序对应于目录名称。
class_names = train_ds.class_names
print(class_names)
得到结果：
['cloudy', 'rain', 'shine', 'sunrise']
2.可视化数据
plt.figure(figsize=(20, 10))
for images, labels in train_ds.take(1):
for i in range(20):
ax = plt.subplot(5, 10, i + 1)
plt.imshow(images[i].numpy().astype("uint8"))
plt.title(class_names[labels[i]])
plt.axis("off")
3.再次检查数据
for image_batch, labels_batch in train_ds:
print(image_batch.shape)
print(labels_batch.shape)
break
得到结果
(32, 180, 180, 3)
(32,)
Image_batch是形状的张量（32,180,180,3）。这是一批形状180x180x3的32张图片（最后一维指的是彩色通道RGB）。
Label_batch是形状（32，）的张量，这些标签对应32张图片。
- 配置数据集
运行
AUTOTUNE = tf.data.AUTOTUNE
train_ds = train_ds.cache().shuffle(1000).prefetch(buffer_size=AUTOTUNE)
val_ds = val_ds.cache().prefetch(buffer_size=AUTOTUNE)
第三部分，构建CNN网络
关于卷积核的计算不懂的可以参考文章：https://blog.csdn.net/qq_38251616/article/details/114278995
layers.Dropout(0.4) 作用是防止过拟合，提高模型的泛化能力。
在上一篇文章花朵识别中，训练准确率与验证准确率相差巨大就是由于模型过拟合导致的
num_classes = 4
model = models.Sequential([
layers.experimental.preprocessing.Rescaling(1./255, input_shape=(img_height, img_width, 3)),
layers.Conv2D(16, (3, 3), activation='relu', input_shape=(img_height, img_width, 3)), # 卷积层1，卷积核3*3
layers.AveragePooling2D((2, 2)), # 池化层1，2*2采样
layers.Conv2D(32, (3, 3), activation='relu'), # 卷积层2，卷积核3*3
layers.AveragePooling2D((2, 2)), # 池化层2，2*2采样
layers.Conv2D(64, (3, 3), activation='relu'), # 卷积层3，卷积核3*3
layers.Dropout(0.3),
layers.Flatten(), # Flatten层，连接卷积层与全连接层
layers.Dense(128, activation='relu'), # 全连接层，特征进一步提取
layers.Dense(num_classes) # 输出层，输出预期结果
])
model.summary() # 打印网络结构
得到结果如下：
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
rescaling_1 (Rescaling) (None, 180, 180, 3) 0
_________________________________________________________________
conv2d_3 (Conv2D) (None, 178, 178, 16) 448
_________________________________________________________________
average_pooling2d_2 (Average (None, 89, 89, 16) 0
_________________________________________________________________
conv2d_4 (Conv2D) (None, 87, 87, 32) 4640
_________________________________________________________________
average_pooling2d_3 (Average (None, 43, 43, 32) 0
_________________________________________________________________
conv2d_5 (Conv2D) (None, 41, 41, 64) 18496
_________________________________________________________________
dropout_1 (Dropout) (None, 41, 41, 64) 0
_________________________________________________________________
flatten_1 (Flatten) (None, 107584) 0
_________________________________________________________________
dense_1 (Dense) (None, 128) 13770880
_________________________________________________________________
dense_2 (Dense) (None, 4) 516
=================================================================
Total params: 13,794,980
Trainable params: 13,794,980
Non-trainable params: 0
第四部分：编译
在准备对模型进行训练之前，还需要再对其进行一些设置。以下内容是在模型的编译步骤中添加的：
损失函数（loss）：用于衡量模型在训练期间的准确率。
优化器（optimizer）：决定模型如何根据其看到的数据和自身的损失函数进行更新。
指标（metrics）：用于监控训练和测试步骤。以下示例使用了准确率，即被正确分类的图像的比率。
opt = tf.keras.optimizers.Adam(learning_rate=0.001)
model.compile(optimizer=opt,
loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
metrics=['accuracy'])
五，训练模型
epochs = 10
history = model.fit(
train_ds,
validation_data=val_ds,
epochs=epochs
)
得到结果如下:
Epoch 1/10
29/29 [==============================] - 11s 234ms/step - loss: 1.0135 - accuracy: 0.6278 - val_loss: 0.6895 - val_accuracy: 0.7244
Epoch 2/10
29/29 [==============================] - 6s 217ms/step - loss: 0.5197 - accuracy: 0.8300 - val_loss: 0.5395 - val_accuracy: 0.8000
Epoch 3/10
29/29 [==============================] - 6s 196ms/step - loss: 0.3535 - accuracy: 0.8700 - val_loss: 0.7650 - val_accuracy: 0.7378
Epoch 4/10
29/29 [==============================] - 6s 203ms/step - loss: 0.2933 - accuracy: 0.9000 - val_loss: 0.4826 - val_accuracy: 0.8133
Epoch 5/10
29/29 [==============================] - 6s 203ms/step - loss: 0.1891 - accuracy: 0.9311 - val_loss: 0.3608 - val_accuracy: 0.8711
Epoch 6/10
29/29 [==============================] - 6s 202ms/step - loss: 0.1985 - accuracy: 0.9189 - val_loss: 0.4878 - val_accuracy: 0.8311
Epoch 7/10
29/29 [==============================] - 6s 197ms/step - loss: 0.1221 - accuracy: 0.9556 - val_loss: 0.3838 - val_accuracy: 0.8800
Epoch 8/10
29/29 [==============================] - 6s 198ms/step - loss: 0.1085 - accuracy: 0.9589 - val_loss: 0.5011 - val_accuracy: 0.8444
Epoch 9/10
29/29 [==============================] - 6s 202ms/step - loss: 0.0704 - accuracy: 0.9789 - val_loss: 0.3517 - val_accuracy: 0.8978
Epoch 10/10
29/29 [==============================] - 6s 204ms/step - loss: 0.1002 - accuracy: 0.9700 - val_loss: 0.4030 - val_accuracy: 0.8800
六，模型评估
acc = history.history['accuracy']
val_acc = history.history['val_accuracy']
loss = history.history['loss']
val_loss = history.history['val_loss']
epochs_range = range(epochs)
plt.figure(figsize=(12, 4))
plt.subplot(1, 2, 1)
plt.plot(epochs_range, acc, label='Training Accuracy')
plt.plot(epochs_range, val_acc, label='Validation Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')
plt.subplot(1, 2, 2)
plt.plot(epochs_range, loss, label='Training Loss')
plt.plot(epochs_range, val_loss, label='Validation Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()
CSDN21天学习挑战赛
https://marketing.csdn.net/p/bdabfb52c5d56532133df2adc1a728fd
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
