深度学习问题记录：Initialization

He initialization

Zero initialization

axes

plt.title("Model with Zeros initialization")
axes = plt.gca()
axes.set_xlim([-1.5,1.5])
axes.set_ylim([-1.5,1.5])
plot_decision_boundary(lambda x: predict_dec(parameters, x.T), train_X, train_Y)

inf

If you see "inf" as the cost after the iteration 0, this is because of numerical roundoff(数值四舍五入); a more numerically sophisticated implementation would fix this. But this isn't worth worrying about for our purposes.

随机初始化过大的缺点

He initialization

结论

relu用He初始化，言外之意似乎是对于sigmoid和than用np.random.randn（）*0.01就可以了？?？