pytorch搭建卷积神经网络【第十五课_激活函数】

---

前言

案例代码https://github.com/2012Netsky/pytorch_cnn/blob/main/4_time_series_bikes.ipynb

#!/usr/bin/env python
# coding: utf-8

# In[1]:


get_ipython().run_line_magic('matplotlib', 'inline')
import numpy as np
import matplotlib.pyplot as plt

import torch
import torch.optim as optim
import torch.nn as nn

torch.set_printoptions(edgeitems=2, linewidth=75)


# In[3]:


input_t = torch.arange(-3, 3.1, 0.1)
print(input_t.shape)
print(input_t)


# In[3]:


activation_list = [
    nn.Tanh(),
    nn.Hardtanh(),
    nn.Sigmoid(),
    
    nn.Softplus(),
    nn.ReLU(),
    nn.LeakyReLU(negative_slope=0.1),

    
    #nn.Tanhshrink(),
    #nn.Softshrink(),
    #nn.Hardshrink(),
]

fig = plt.figure(figsize=(14, 28), dpi=600)

for i, activation_func in enumerate(activation_list):
    subplot = fig.add_subplot(len(activation_list), 3, i+1)
    subplot.set_title(type(activation_func).__name__)
    
    output_t = activation_func(input_t)
    
    plt.grid()
    plt.plot(input_t.numpy(), input_t.numpy(),'k', linewidth=1)
    plt.plot([-3,3], [0,0], 'k', linewidth=1)
    plt.plot([0,0], [-3,3], 'k', linewidth=1)
    plt.plot(input_t.numpy(), output_t.numpy(), 'r', linewidth=3)

output_t


# In[4]:


a = lambda x: nn.Tanh()(-2 * x - 1.25)
b = lambda x: nn.Tanh()( 1 * x + 0.75)
c = lambda x: nn.Tanh()( 4 * x + 1.)
d = lambda x: nn.Tanh()(-3 * x + 1.5)

wb_list = [
    ('A: tanh(-2 * x - 1.25)', a(input_t)),
    ('B: tanh(1 * x + 0.75)', b(input_t)),
    ('A + B', a(input_t) + b(input_t)),

    ('C: tanh(4 * x + 1.0)', c(input_t)),
    ('D: tanh(-3 * x - 1.5)', d(input_t)),
    ('C + D', c(input_t) + d(input_t)),

    ('C(A + B)', c(a(input_t) + b(input_t))),
    ('D(A + B)', d(a(input_t) + b(input_t))),
     
    ('C(A+B) + D(A+B)', 
     c(a(input_t) + b(input_t)) + d(a(input_t) + b(input_t))),
]

fig = plt.figure(figsize=(14, 42), dpi=600)

for i, (title_str, output_t) in enumerate(wb_list):
    subplot = fig.add_subplot(len(wb_list), 3, i+1)
    subplot.set_title(title_str)
    
    plt.grid()
    plt.plot(input_t.numpy(), input_t.numpy(),'k', linewidth=1)
    plt.plot([-3,3], [0,0], 'k', linewidth=1)
    plt.plot([0,0], [-3,3], 'k', linewidth=1)
    plt.plot(input_t.numpy(), output_t.numpy(), 'r', linewidth=3)

output_t


# In[5]:


fig = plt.figure(figsize=(14, 14), dpi=600)

subplot = fig.add_subplot(1, 1, 1)
#subplot.set_title(type(activation_func).__name__)

output_t = nn.Tanh()(input_t)

#plt.grid()
plt.plot(input_t.numpy(), input_t.numpy(),'k', linewidth=1)
plt.plot([-3,3], [0,0], 'k', linewidth=1)
plt.plot([0,0], [-3,3], 'k', linewidth=1)
plt.plot(input_t.numpy(), output_t.numpy(), 'r', linewidth=3)

output_t


# In[ ]:

总结