神经网络学习ING-莫烦python-RNN回归-程序理解

RNN：保留上一个输入的学习状态，输出不仅与当前输入有关，还与学习状态有关

import torch
from torch import nn
import numpy as np
import matplotlib.pyplot as plt

# torch.manual_seed(1)    # reproducible

# Hyper Parameters
TIME_STEP = 10  # rnn time step;时间点
INPUT_SIZE = 1  # rnn input size；每个时间点输入数据数量
LR = 0.02  # learning rate

# show data，这里的数据仅仅是用来画图展示sin和cos图像的，与RNN的输入数据无关
# steps = np.linspace(0, np.pi * 2, 100, dtype=np.float32)  # float32 for converting torch FloatTensor
# x_np = np.sin(steps)
# y_np = np.cos(steps)
# plt.figure(num='图像一')
# plt.plot(steps, y_np, 'r-', label='target (cos)')
# plt.plot(steps, x_np, 'b-', label='input (sin)')
# plt.legend(loc='best')
#同时显示两个图像需要使用figure

#构建网络
class RNN(nn.Module):
    def __init__(self):
        super(RNN, self).__init__()

        self.rnn = nn.RNN(
            input_size=INPUT_SIZE,
            hidden_size=32,  # rnn hidden unit
            num_layers=1,  # number of rnn layer
            batch_first=True,  # input & output will has batch size as 1s dimension. e.g. (batch, time_step, input_size)
        )
        #batch_first 默认为Flase，即在RNN中batch_size放在第二维，便于并行计算
        self.out = nn.Linear(32, 1)

    def forward(self, x, h_state):
        # x (batch, time_step, input_size)
        # h_state (n_layers, batch, hidden_size)
        # r_out (batch, time_step, hidden_size)
        r_out, h_state = self.rnn(x, h_state)
        #print(r_out[0].size()) #torch.Size([10,32])
        #每一个batch是一个10*32的矩阵

    ###############
        #这里的r_out是rnn出来的输出，size为(1,10,32)
        #需要把每一个time_step的r_out送到self.out中全连接

        outs = []  # save all predictions
        for time_step in range(r_out.size(1)):  # calculate output for each time step
            outs.append(self.out(r_out[:, time_step, :]))
        return torch.stack(outs, dim=1), h_state
        #将outs在新的维度上叠加，并且把这个新的维度作为第2维
        # instead, for simplicity, you can replace above codes by follows
        # r_out = r_out.view(-1, 32)
        # outs = self.out(r_out)
        # outs = outs.view(-1, TIME_STEP, 1)
        # return outs, h_state

        # or even simpler, since nn.Linear can accept inputs of any dimension
        # and returns outputs with same dimension except for the last
        # outs = self.out(r_out)
        # return outs
    #############

rnn = RNN()
print(rnn)

optimizer = torch.optim.Adam(rnn.parameters(), lr=LR)  # optimize all cnn parameters
loss_func = nn.MSELoss()

h_state = None  # for initial hidden state

plt.figure(num='图像二', figsize=(12, 5))
plt.ion()  # continuously plot

for step in range(100):
    start, end = step * np.pi, (step + 1) * np.pi  # time range
    # use sin predicts cos
    steps = np.linspace(start, end, TIME_STEP, dtype=np.float32,
                        endpoint=False)  # float32 for converting torch FloatTensor
    #从0-100pi，每个pi间隔取10个点，共1000个数据点
    #计算函数值
    x_np = np.sin(steps)
    y_np = np.cos(steps)
    #将函数值包装成RNN输入数据的size(batch, time_step, input_size)
    x = torch.from_numpy(x_np[np.newaxis, :, np.newaxis])  # shape (batch, time_step, input_size)
    y = torch.from_numpy(y_np[np.newaxis, :, np.newaxis])
    #print(y.size())
    prediction, h_state = rnn(x, h_state)  # rnn output
    #print(prediction)
    #这里的prediction是经过全连接之后的输出，size为[1,10,1]
    # !! next step is important !!
    h_state = h_state.data  # repack the hidden state, break the connection from last iteration

    loss = loss_func(prediction, y)  # calculate loss;训练X数据，预测Y值
    optimizer.zero_grad()  # clear gradients for this training step
    loss.backward()  # backpropagation, compute gradients
    optimizer.step()  # apply gradients

    # plotting
    plt.plot(steps, y_np.flatten(), 'r-') #y的真实值
    plt.plot(steps, prediction.data.numpy().flatten(), 'b-') #y的预测值
    plt.draw();
    plt.pause(0.05)

plt.ioff()
plt.show()

原程序来自：
https://github.com/MorvanZhou/PyTorch-Tutorial