第一种：分开来画 import matplotlib.pyplot as plt import numpy as np plt.rcParams['axes.unicode_minus'] = False def sigmoid(x): return 1.0/(1.0+np.exp(-x)) def tanh(x): return (np.exp(x) - np.exp(-x))/(np.exp(x)+np.exp(-x)) x = np.linspace(-8,8) fig =