lstm - pytorch

1 lstm做分类

2 lstm做回归

3 lstm的参数输入

 

1 图像分类任务：

train_loader中取出来的变量维度： torch.Size([64, 1, 28, 28])

需要将其改成：torch.Size([64, 28, 28]) 代表batchsize , height, weight 

lstm的参数设置：

input: 形状的输入（seq_len，batch，input_size）

h_0 :形状（num_layers * num_directions，batch，hidden_​​size）：张量，包含批次中每个元素的初始隐藏状态。如果LSTM是双向的，则num_directions应该为2，否则应为1。

c_0 :形状（num_layers * num_directions，batch，hidden_​​size）：张量，包含批次中每个元素的初始单元状态。

如果未提供（h_0，c_0），则h_0和c_0均默认为零。

batch_first :如果为True，则输入和输出张量按（batch，seq，feature）提供

output：形状的输出（seq_len，batch，num_directions * hidden_​​size）：张量，包含每个t的LSTM最后一层的输出特征（h_t）。如果已将a 用作输入，则输出也将是打包序列

import torch
from torch import nn
import torchvision.datasets as dsets
import torchvision.transforms as transforms
import matplotlib.pyplot as plt
import os

# torch.manual_seed(1)    # reproducible

EPOCH = 1               # train the training data n times, to save time, we just train 1 epoch
BATCH_SIZE = 64
TIME_STEP = 28          # rnn time step / image height
INPUT_SIZE = 28         # rnn input size / image width
LR = 0.01               # learning rate
DOWNLOAD_MNIST = False
if not(os.path.exists('./mnist/')) or not os.listdir('./mnist/'):
    # not mnist dir or mnist is empyt dir
    DOWNLOAD_MNIST = True

# Mnist digital dataset
train_data = dsets.MNIST(
    root='./mnist/',
    train=True,                         # this is training data
    transform=transforms.ToTensor(),    # Converts a PIL.Image or numpy.ndarray to
                                        # torch.FloatTensor of shape (C x H x W) and normalize in the range [0.0, 1.0]
    download=DOWNLOAD_MNIST,            # download it if you don't have it
)

# plot one example
print(train_data.train_data.size())     # (60000, 28, 28)
print(train_data.train_labels.size())   # (60000)
plt.imshow(train_data.train_data[0].numpy(), cmap='gray')
plt.title('%i' % train_data.train_labels[0])
plt.show()

# Data Loader for easy mini-batch return in training
train_loader = torch.utils.data.DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True)

# convert test data into Variable, pick 2000 samples to speed up testing
test_data = dsets.MNIST(root='./mnist/', train=False, transform=transforms.ToTensor())
test_x = test_data.test_data.type(torch.FloatTensor)[:2000]/255.   # shape (2000, 28, 28) value in range(0,1)
test_y = test_data.test_labels.numpy()[:2000]    # covert to numpy array

input = torch.randn(5, 30, 100)   # seq_len，batch，input_size
h0 = torch.randn(4, 30, 200)      # num_layers * num_directions，batch，hidden_​​size
c0 = torch.randn(4, 30, 200)      # num_layers * num_directions，batch，hidden_​​size
rnn = nn.LSTM(100, 200, 4)         # input_size , hidden_​​size ,num_layers
output, (hn, cn) = rnn(input, (h0, c0)) #seq_len，batch，num_directions * hidden_​​size
print(output.shape)

class RNN(nn.Module):
    def __init__(self):
        super(RNN, self).__init__()
        self.rnn = nn.LSTM(         # if use nn.RNN(), it hardly learns
            input_size=28,
            hidden_size=64,         # rnn hidden unit
            num_layers=3,           # number of rnn layer
            batch_first=True,       # input & output will has batch size as 1s dimension. e.g. (batch, time_step, input_size)
        )
        self.out = nn.Linear(64, 10)
    def forward(self, x):
        # x shape (batch, time_step, input_size)
        # r_out shape (batch, time_step, output_size)
        # h_n shape (n_layers, batch, hidden_size)
        # h_c shape (n_layers, batch, hidden_size)
        print(x.shape)
        r_out, (h_n, h_c) = self.rnn(x, None)   # None represents zero initial hidden state
        print(r_out.shape)
        # choose r_out at the last time step
        out = self.out(r_out[:, -1, :])
        return out


rnn = RNN()
print(rnn)

optimizer = torch.optim.Adam(rnn.parameters(), lr=LR)   # optimize all cnn parameters
loss_func = nn.CrossEntropyLoss()                       # the target label is not one-hotted

# training and testing
for epoch in range(EPOCH):
    for step, (b_x, b_y) in enumerate(train_loader):        # gives batch data
        b_x = b_x.view(-1, 28, 28)              # reshape x to (batch, time_step, input_size)

        output = rnn(b_x)                               # rnn output
        loss = loss_func(output, b_y)                   # cross entropy loss
        optimizer.zero_grad()                           # clear gradients for this training step
        loss.backward()                                 # backpropagation, compute gradients
        optimizer.step()                                # apply gradients

        if step % 50 == 0:
            test_output = rnn(test_x)                   # (samples, time_step, input_size)
            pred_y = torch.max(test_output, 1)[1].data.numpy()
            accuracy = float((pred_y == test_y).astype(int).sum()) / float(test_y.size)
            print('Epoch: ', epoch, '| train loss: %.4f' % loss.data.numpy(), '| test accuracy: %.2f' % accuracy)

# print 10 predictions from test data
test_output = rnn(test_x[:10].view(-1, 28, 28))
pred_y = torch.max(test_output, 1)[1].data.numpy()
print(pred_y, 'prediction number')
print(test_y[:10], 'real number')

2 Lstm回归

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

TIME_STEP = 10      # rnn time step
INPUT_SIZE = 1      # rnn input size
LR = 0.02           # learning rate

# show data
# 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.plot(steps, y_np, 'r-', label='target (cos)')
# plt.plot(steps, x_np, 'b-', label='input (sin)')
# plt.legend(loc='best')
# plt.show()


class RNN(nn.Module):
    def __init__(self):
        super(RNN, self).__init__()

        self.rnn = nn.RNN(
            input_size=1,
            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)
        )
        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)     # output: torch.Size([1, 10, 32])

        # print('r_out:' , r_out.shape)
        # outs = []    # save all predictions
        # L1= r_out.size(1)
        # for time_step in range(L1):    # calculate output for each time step
        #     tmp1 = r_out[:, time_step, :]
        #     tmp2 = self.out(tmp1)
        #     outs.append(tmp2)
        #
        # outx = torch.stack(outs, dim=1)
        # return outx , h_state

        # 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(1, figsize=(12, 5))
# plt.ion()           # continuously plot

for step in range(100):
    start, end = step * np.pi, (step+1)*np.pi   # time range
    steps = np.linspace(start, end, TIME_STEP, dtype=np.float32, endpoint=False)  # float32 for converting torch FloatTensor
    x_np = np.sin(steps)
    y_np = np.cos(steps)

    x = torch.from_numpy(x_np[np.newaxis, :, np.newaxis])    # shape (batch, time_step, input_size)
    # print('x:',x.shape)
    y = torch.from_numpy(y_np[np.newaxis, :, np.newaxis])
    # print('y:',y.shape)
    prediction, h_state = rnn(x, h_state)   # rnn output
    # print(prediction.shape)
    # !! 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
    optimizer.zero_grad()                   # clear gradients for this training step
    loss.backward()                         # backpropagation, compute gradients
    optimizer.step()                        # apply gradients

    print('step: ',  step, '| train loss: %.4f' % loss.data.numpy(), '| test accuracy: %.2f' % 1)

    # plotting
    # plt.plot(steps, y_np.flatten(), 'r-')
    # plt.plot(steps, prediction.data.numpy().flatten(), 'b-')
    # plt.draw(); plt.pause(0.05)

plt.ioff()
plt.show()