当前位置：首页 > article >正文

深度学习作业九 RNN-SRN-Seq2Seq

article 2024/11/29 23:06:13

1. 实现SRN

（1）使用Numpy

（2）在1的基础上，增加激活函数tanh

（3）使用nn.RNNCell实现

（4）使用nn.RNN实现

2. 实现“序列到序列”

3. “编码器-解码器”的简单实现

4.谈一谈对“序列”、“序列到序列”的理解

4.1序列

4.2序列到序列

5.总结本周理论课和作业，写心得体会

1. 实现SRN

（1）使用Numpy

代码如下：

import numpy as np

inputs = np.array([[1., 1.],
                   [1., 1.],
                   [2., 2.]])  # 初始化输入序列
print('inputs is ', inputs)

state_t = np.zeros(2, )  # 初始化存储器
print('state_t is ', state_t)

w1, w2, w3, w4, w5, w6, w7, w8 = 1., 1., 1., 1., 1., 1., 1., 1.
U1, U2, U3, U4 = 1., 1., 1., 1.
for input_t in inputs:
    print('inputs is ', input_t)
    print('state_t is ', state_t)
    in_h1 = np.dot([w1, w3], input_t) + np.dot([U2, U4], state_t)
    in_h2 = np.dot([w2, w4], input_t) + np.dot([U1, U3], state_t)
    state_t = in_h1, in_h2
    output_y1 = np.dot([w5, w7], [in_h1, in_h2])
    output_y2 = np.dot([w6, w8], [in_h1, in_h2])
    print('output_y is ', output_y1, output_y2)
    print('---------------------')

运行结果：

如何算的呢？

如图，第一次输入x1，x2,均为1，此时a1，a2值为0，权重均为1，所以h1=w1*x1+w3*x2=2，h2=w2*x1+w4*x2=2，同时将h1，h2存入a1，a2中，a1=2，a2=2，之后y1=w5*h1+w7*h2=4，y2=w6*h1+w8*h2=4.

第二次输入x1，x2均为1，此时a1=2，a2=2，所以h1=w1*x1+w3*x2+u2*a1+u4*a2=6，h2=w2*x1+w4*x2+u1*a1+u3*a2=6，同时将h1，h2存入a1，a2中，a1=6，a2=6，之后y1=w5*h1+w7*h2=12，y2=w6*h1+w8*h2=12.

第三次输入x1，x2均为2，此时a1=6，a2=6，所以h1=w1*x1+w3*x2+u2*a1+u4*a2=16，h2=w2*x1+w4*x2+u1*a1+u3*a2=16，同时将h1，h2存入a1，a2中，a1=16，a2=16，之后y1=w5*h1+w7*h2=32，y2=w6*h1+w8*h2=32.

（2）在1的基础上，增加激活函数tanh

注：只在输入层到隐藏层使用了激活函数tanh，隐藏层到输出层激活函数依旧为linear

代码如下：

for input_t in inputs:
    print('inputs is ', input_t)
    print('state_t is ', state_t)
    in_h1 = np.tanh(np.dot([w1, w3], input_t) + np.dot([U2, U4], state_t))
    in_h2 = np.tanh(np.dot([w2, w4], input_t) + np.dot([U1, U3], state_t))
    state_t = in_h1, in_h2
    output_y1 = np.dot([w5, w7], [in_h1, in_h2])
    output_y2 = np.dot([w6, w8], [in_h1, in_h2])
    print('output_y is ', output_y1, output_y2)
    print('---------------------------------------------')

运行结果：

（3）使用nn.RNNCell实现

nn.RNNCell 是一个单步 RNN 单元，用于手动处理每一个时间步的数据。它适合需要细粒度控制或对序列操作的情景。

参数

input_size: 输入特征的维度。
hidden_size: 隐藏状态的维度。
bias: 是否包含偏置项，默认是 True。
nonlinearity: 激活函数类型，可选 'tanh' 或 'relu'，默认是 'tanh'。

使用步骤

定义一个 nn.RNNCell 实例，指定输入和隐藏层的维度。
初始化隐藏状态，一般是零张量。
手动遍历输入序列，在每个时间步调用 rnn_cell(input_t, hidden_t)。

主要就是实现上图这个功能（store）

代码如下：

batch_size = 1
seq_len = 3  # 序列长度
input_size = 2  # 输入序列维度
hidden_size = 2  # 隐藏层维度
output_size = 2  # 输出层维度

# RNNCell
cell = torch.nn.RNNCell(input_size=input_size, hidden_size=hidden_size)
# 初始化参数，权重为1，偏置为0
for name, param in cell.named_parameters():
    if name.startswith("weight"):
        torch.nn.init.ones_(param)
    else:
        torch.nn.init.zeros_(param)
# 线性层
liner = torch.nn.Linear(hidden_size, output_size)
liner.weight.data = torch.Tensor([[1, 1], [1, 1]])
liner.bias.data = torch.Tensor([0.0])

seq = torch.Tensor([[[1, 1]],
                    [[1, 1]],
                    [[2, 2]]])
hidden = torch.zeros(batch_size, hidden_size)
output = torch.zeros(batch_size, output_size)

for idx, input in enumerate(seq):
    print('=' * 20, idx, '=' * 20)

    print('Input :', input)
    print('hidden :', hidden)

    hidden = cell(input, hidden)
    output = liner(hidden)
    print('output :', output)

运行结果：

对比上面numpy实现的结果相差不大

（4）使用nn.RNN实现

nn.RNN 是一个完整的 RNN 模块，可以一次性处理整个序列。它内部会自动迭代时间步并更新隐藏状态。

参数

input_size: 输入特征的维度。
hidden_size: 隐藏状态的维度。
num_layers: RNN 的层数，默认为 1。
nonlinearity: 激活函数类型，可选 'tanh' 或 'relu'，默认是 'tanh'。
batch_first: 若为 True，输入和输出的维度格式为 (batch_size, seq_len, input_size)；否则为 (seq_len, batch_size, input_size)。

返回值

output: 所有时间步的隐藏状态 (seq_len, batch_size, hidden_size)。
h_n: 最后一个时间步的隐藏状态 (num_layers, batch_size, hidden_size)。

使用步骤

定义一个 nn.RNN 实例，指定输入和隐藏层的维度。
初始化隐藏状态（可选）。
将输入序列和初始隐藏状态传入 rnn(inputs, hidden)。

代码如下：

batch_size = 1
seq_len = 3
input_size = 2
hidden_size = 2
num_layers = 1
output_size = 2

cell = torch.nn.RNN(input_size=input_size, hidden_size=hidden_size, num_layers=num_layers)
for name, param in cell.named_parameters():  # 初始化参数
    if name.startswith("weight"):
        torch.nn.init.ones_(param)
    else:
        torch.nn.init.zeros_(param)

# 线性层
liner = torch.nn.Linear(hidden_size, output_size)
liner.weight.data = torch.Tensor([[1, 1], [1, 1]])
liner.bias.data = torch.Tensor([0.0])

inputs = torch.Tensor([[[1, 1]],
                       [[1, 1]],
                       [[2, 2]]])
hidden = torch.zeros(num_layers, batch_size, hidden_size)
out, hidden = cell(inputs, hidden)
print("out:", out)
print("hidden:", hidden)
print('--------------------------------------')

print('Input :', inputs[0])
print('hidden:', 0, 0)
print('Output:', liner(out[0]))
print('--------------------------------------')
print('Input :', inputs[1])
print('hidden:', out[0])
print('Output:', liner(out[1]))
print('--------------------------------------')
print('Input :', inputs[2])
print('hidden:', out[1])
print('Output:', liner(out[2]))

运行结果：

结果对比一样。

2. 实现“序列到序列”

观看视频，学习RNN原理，并实现视频P12中的教学案例

12.循环神经网络（基础篇）_哔哩哔哩_bilibili

代码如下：

input_size = 4
hidden_size = 4
batch_size = 1

idx2char = ['e', 'h', 'l', 'o']
x_data = [1, 0, 2, 2, 3]
y_data = [3, 1, 2, 3, 2]

one_hot_lookup = [[1, 0, 0, 0],
                  [0, 1, 0, 0],
                  [0, 0, 1, 0],
                  [0, 0, 0, 1]]
x_one_hot = [one_hot_lookup[x] for x in x_data]

inputs = torch.Tensor(x_one_hot).view(-1, batch_size, input_size)
labels = torch.LongTensor(y_data).view(-1, 1)


class Model(torch.nn.Module):
    def __init__(self, input_size, hidden_size, batch_size):
        super(Model, self).__init__()
        self.batch_size = batch_size
        self.input_size = input_size
        self.hidden_size = hidden_size
        self.rnncell = torch.nn.RNNCell(input_size=self.input_size, hidden_size=self.hidden_size)

    def forward(self, input, hidden):
        hidden = self.rnncell(input, hidden)
        return hidden

    def init_hidden(self):
        return torch.zeros(self.batch_size, self.hidden_size)


net = Model(input_size, hidden_size, batch_size)
criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(net.parameters(), lr=0.1)

for epoch in range(15):
    loss = 0
    optimizer.zero_grad()
    hidden = net.init_hidden()
    print('Predicted string: ', end='')
    for input, label in zip(inputs, labels):
        hidden = net(input, hidden)
        loss += criterion(hidden, label)
        _, idx = hidden.max(dim=1)
        print(idx2char[idx.item()], end='')
    loss.backward()
    optimizer.step()
    print(', Epoch [%d/15] loss=%.4f' % (epoch + 1, loss.item()))

运行结果：

3. “编码器-解码器”的简单实现

参考链接：Seq2Seq的PyTorch实现 - mathor

代码如下：

import torch
import numpy as np
import torch.nn as nn
import torch.utils.data as Data

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

letter = [c for c in 'SE?abcdefghijklmnopqrstuvwxyz']
letter2idx = {n: i for i, n in enumerate(letter)}

seq_data = [['man', 'women'], ['black', 'white'], ['king', 'queen'], ['girl', 'boy'], ['up', 'down'], ['high', 'low']]

# Seq2Seq Parameter
n_step = max([max(len(i), len(j)) for i, j in seq_data])  # max_len(=5)
n_hidden = 128
n_class = len(letter2idx)  # classfication problem
batch_size = 3


def make_data(seq_data):
    enc_input_all, dec_input_all, dec_output_all = [], [], []

    for seq in seq_data:
        for i in range(2):
            seq[i] = seq[i] + '?' * (n_step - len(seq[i]))  # 'man??', 'women'

        enc_input = [letter2idx[n] for n in (seq[0] + 'E')]  # ['m', 'a', 'n', '?', '?', 'E']
        dec_input = [letter2idx[n] for n in ('S' + seq[1])]  # ['S', 'w', 'o', 'm', 'e', 'n']
        dec_output = [letter2idx[n] for n in (seq[1] + 'E')]  # ['w', 'o', 'm', 'e', 'n', 'E']

        enc_input_all.append(np.eye(n_class)[enc_input])
        dec_input_all.append(np.eye(n_class)[dec_input])
        dec_output_all.append(dec_output)  # not one-hot

    # make tensor
    return torch.Tensor(enc_input_all), torch.Tensor(dec_input_all), torch.LongTensor(dec_output_all)


enc_input_all, dec_input_all, dec_output_all = make_data(seq_data)


class TranslateDataSet(Data.Dataset):
    def __init__(self, enc_input_all, dec_input_all, dec_output_all):
        self.enc_input_all = enc_input_all
        self.dec_input_all = dec_input_all
        self.dec_output_all = dec_output_all

    def __len__(self):  # return dataset size
        return len(self.enc_input_all)

    def __getitem__(self, idx):
        return self.enc_input_all[idx], self.dec_input_all[idx], self.dec_output_all[idx]


loader = Data.DataLoader(TranslateDataSet(enc_input_all, dec_input_all, dec_output_all), batch_size, True)


# Model
class Seq2Seq(nn.Module):
    def __init__(self):
        super(Seq2Seq, self).__init__()
        self.encoder = nn.RNN(input_size=n_class, hidden_size=n_hidden, dropout=0.5)  # encoder
        self.decoder = nn.RNN(input_size=n_class, hidden_size=n_hidden, dropout=0.5)  # decoder
        self.fc = nn.Linear(n_hidden, n_class)

    def forward(self, enc_input, enc_hidden, dec_input):
        enc_input = enc_input.transpose(0, 1)  # enc_input: [n_step+1, batch_size, n_class]
        dec_input = dec_input.transpose(0, 1)  # dec_input: [n_step+1, batch_size, n_class]

        # h_t : [num_layers(=1) * num_directions(=1), batch_size, n_hidden]
        _, h_t = self.encoder(enc_input, enc_hidden)
        # outputs : [n_step+1, batch_size, num_directions(=1) * n_hidden(=128)]
        outputs, _ = self.decoder(dec_input, h_t)

        model = self.fc(outputs)  # model : [n_step+1, batch_size, n_class]
        return model


model = Seq2Seq().to(device)
criterion = nn.CrossEntropyLoss().to(device)
optimizer = torch.optim.Adam(model.parameters(), lr=0.001)

for epoch in range(5000):
    for enc_input_batch, dec_input_batch, dec_output_batch in loader:
        # make hidden shape [num_layers * num_directions, batch_size, n_hidden]
        h_0 = torch.zeros(1, batch_size, n_hidden).to(device)

        (enc_input_batch, dec_intput_batch, dec_output_batch) = (
            enc_input_batch.to(device), dec_input_batch.to(device), dec_output_batch.to(device))
        # enc_input_batch : [batch_size, n_step+1, n_class]
        # dec_intput_batch : [batch_size, n_step+1, n_class]
        # dec_output_batch : [batch_size, n_step+1], not one-hot
        pred = model(enc_input_batch, h_0, dec_intput_batch)
        # pred : [n_step+1, batch_size, n_class]
        pred = pred.transpose(0, 1)  # [batch_size, n_step+1(=6), n_class]
        loss = 0
        for i in range(len(dec_output_batch)):
            # pred[i] : [n_step+1, n_class]
            # dec_output_batch[i] : [n_step+1]
            loss += criterion(pred[i], dec_output_batch[i])
        if (epoch + 1) % 1000 == 0:
            print('Epoch:', '%04d' % (epoch + 1), 'cost =', '{:.6f}'.format(loss))

        optimizer.zero_grad()
        loss.backward()
        optimizer.step()


# Test
def translate(word):
    enc_input, dec_input, _ = make_data([[word, '?' * n_step]])
    enc_input, dec_input = enc_input.to(device), dec_input.to(device)
    # make hidden shape [num_layers * num_directions, batch_size, n_hidden]
    hidden = torch.zeros(1, 1, n_hidden).to(device)
    output = model(enc_input, hidden, dec_input)
    # output : [n_step+1, batch_size, n_class]

    predict = output.data.max(2, keepdim=True)[1]  # select n_class dimension
    decoded = [letter[i] for i in predict]
    translated = ''.join(decoded[:decoded.index('E')])

    return translated.replace('?', '')


print('test')
print('man ->', translate('man'))
print('mans ->', translate('mans'))
print('king ->', translate('king'))
print('black ->', translate('black'))
print('up ->', translate('up'))
print('old ->', translate('old'))
print('high ->', translate('high'))

运行结果：