Github

• https://github.com/matplotlib/matplotlib

官网

• https://matplotlib.org/stable/

文档

• https://matplotlib.org/stable/api/index.html

简介

matplotlib 是 Python 中最常用的绘图库之一，用于创建各种类型的静态、动态和交互式可视化。以下是对 matplotlib 的详细介绍，包括其安装、基本使用方法、常见绘图类型、以及如何进行高级绘图和自定义图形。

动态显示训练过程中的数据和模型的决策边界

安装

pip install tensorflow==2.13.1
pip install matplotlib==3.7.5
pip install numpy==1.24.3

源码

import numpy as np
import tensorflow as tf
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
import matplotlib.pyplot as plt
from matplotlib.colors import ListedColormap

# 生成数据
np.random.seed(0)
num_samples_per_class = 500
negative_samples = np.random.multivariate_normal(
    mean=[0, 3],
    cov=[[1, 0.5], [0.5, 1]],
    size=num_samples_per_class
)
positive_samples = np.random.multivariate_normal(
    mean=[3, 0],
    cov=[[1, 0.5], [0.5, 1]],
    size=num_samples_per_class
)

inputs = np.vstack((negative_samples, positive_samples)).astype(np.float32)
targets = np.vstack((np.zeros((num_samples_per_class, 1)), np.ones((num_samples_per_class, 1)))).astype(np.float32)

# 将数据分为训练集和测试集
train_size = int(0.8 * len(inputs))
X_train, X_test = inputs[:train_size], inputs[train_size:]
y_train, y_test = targets[:train_size], targets[train_size:]

# 构建二分类模型
model = Sequential([
    # 输入层：输入形状为 (2,)
    # 第一个隐藏层：包含 4 个节点，激活函数使用 ReLU
    Dense(4, activation='relu', input_shape=(2,)),
    
    # 输出层：包含 1 个节点，激活函数使用 Sigmoid（因为是二分类问题）
    Dense(1, activation='sigmoid')
])

# 编译模型
# 指定优化器为 Adam，损失函数为二分类交叉熵，评估指标为准确率
model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])

# 准备绘图
fig, ax = plt.subplots()
cmap_light = ListedColormap(['#FFAAAA', '#AAAAFF'])
cmap_bold = ListedColormap(['#FF0000', '#0000FF'])

# 动态绘制函数
def plot_decision_boundary(epoch, logs):
    ax.clear()
    x_min, x_max = X_train[:, 0].min() - 1, X_train[:, 0].max() + 1
    y_min, y_max = X_train[:, 1].min() - 1, X_train[:, 1].max() + 1
    xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.1),
                         np.arange(y_min, y_max, 0.1))
    grid = np.c_[xx.ravel(), yy.ravel()]
    probs = model.predict(grid).reshape(xx.shape)
    ax.contourf(xx, yy, probs, alpha=0.8, cmap=cmap_light)
    ax.scatter(X_train[:, 0], X_train[:, 1], c=y_train[:, 0], edgecolor='k', cmap=cmap_bold)
    ax.set_title(f'Epoch {epoch+1}')
    plt.draw()
    plt.pause(0.01)

# 自定义回调函数
class PlotCallback(tf.keras.callbacks.Callback):
    def on_epoch_end(self, epoch, logs=None):
        plot_decision_boundary(epoch, logs)

# 训练模型并动态显示
plot_callback = PlotCallback()
model.fit(X_train, y_train, epochs=50, batch_size=16, callbacks=[plot_callback])

# 评估模型
loss, accuracy = model.evaluate(X_test, y_test)
print(f"Test Loss: {loss}")
print(f"Test Accuracy: {accuracy}")

plt.show()