PyTorch: RNN实战详解之生成名字

版权声明：博客文章都是作者辛苦整理的，转载请注明出处，谢谢！ http://blog.csdn.net/m0_37306360/article/details/79316964

介绍

上一篇我们讲了如何在PyTorch框架下用RNN分类名字http://blog.csdn.net/m0_37306360/article/details/79316013，本文讲如何用RNN生成特定语言(类别)的名字。我们使用上一篇同样的数据集。不同的是，不是根据输入的名字来预测此名字是那种语言的(读完名字的所有字母之后，我们不是预测一个类别)。而是一次输入一个类别并输出一个字母。 反复预测字符以生成名字。

PyTorch之RNN实战生成

定义网络

class RNN(nn.Module):
def __init__(self, input_size, hidden_size, output_size):
super(RNN, self).__init__()
self.hidden_size = hidden_size

self.i2h = nn.Linear(n_categories + input_size + hidden_size, hidden_size)
self.i2o = nn.Linear(n_categories + input_size + hidden_size, output_size)
self.o2o = nn.Linear(hidden_size + output_size, output_size)
self.dropout = nn.Dropout(0.1)
self.softmax = nn.LogSoftmax(dim=1)

def forward(self, category, input, hidden):
input_combined = torch.cat((category, input, hidden), 1)
output = self.i2o(input_combined)
hidden = self.i2h(input_combined)
output_combined = torch.cat((output, hidden), 1)
output = self.o2o(output_combined)
output = self.dropout(output)
output = self.softmax(output)
return output, hidden

def initHidden(self):
return Variable(torch.zeros(1, self.hidden_size))

数据预处理

其他的数据预处理和上一篇文章类似。主要的不同是如何构建训练样本对。

对于每个时间步（即训练词中的每个字母），网络的输入是（类别，当前字母，隐藏状态），输出将是（下一个字母，下一个隐藏状态）。 因此，对于每个训练集，我们需要：类别，一组输入字母和一组输出(目标)字母。

我们可以很简单的获取(类别，名字(字母序列))，但是我们如何用这些数据构建输入字母序列和输出字母序列呢？由于我们预测了每个时间步的当前字母的下一个字母，因此字母对是来自行的连续字母的组 - 例如对于序列(“ABCD”)，构建(“A”, B”), (“B”,“C”), (“C”, “D”), (“D”,“EOS”). 如图：

训练网络

与仅使用最后一个输出的分类相比，我们在每一步都进行了预测，因此我们需要计算每一步的损失。

for i in range(input_line_tensor.size()[0]):
output, hidden = rnn(category_tensor, input_line_tensor[i], hidden)
loss += criterion(output, target_line_tensor[i])

采样

为了取样，我们给网络输入一个字母，得到下一个字母，把它作为下一个字母的输入，并重复，直到返回EOS结束符。

整个生成过程如下：

1. 为输入类别，开始字母和空的隐藏状态创建张量

2. 用开始字母创建一个字符串:output_name

3. 达到最大输出长度:

(1).将当前的字母送入网络

(2).从输出获取下一个字母和下一个隐藏状态

(3).如果这个字符是EOS，生成结束

(4).如果是普通字母，添加到output_name并继续

4.返回最后生成的名字

完整代码

from io import open
import glob
import unicodedata
import string

all_letters = string.ascii_letters + " .,;'-"
n_letters = len(all_letters) + 1 # Plus EOS marker

def findFiles(path): return glob.glob(path)

# Turn a Unicode string to plain ASCII, thanks to http://stackoverflow.com/a/518232/2809427 def unicodeToAscii(s):
return ''.join(
c for c in unicodedata.normalize('NFD', s)
if unicodedata.category(c) != 'Mn'
and c in all_letters
)

# Read a file and split into lines
def readLines(filename):
lines = open(filename, encoding='utf-8').read().strip().split('\n')
return [unicodeToAscii(line) for line in lines]

# Build the category_lines dictionary, a list of lines per category
category_lines = {}
all_categories = []
for filename in findFiles('nlpdata/data/names/*.txt'):
category = filename.split('/')[-1].split('.')[0]
category = category.split('\\')[1]
all_categories.append(category)
lines = readLines(filename)
category_lines[category] = lines

n_categories = len(all_categories)

# print('# categories:', n_categories, all_categories)
# print(unicodeToAscii("O'Néàl"))

import torch
import torch.nn as nn
from torch.autograd import Variable

class RNN(nn.Module):
def __init__(self, input_size, hidden_size, output_size):
super(RNN, self).__init__()
self.hidden_size = hidden_size
self.i2h = nn.Linear(n_categories + input_size + hidden_size, hidden_size)
self.i2o = nn.Linear(n_categories + input_size + hidden_size, output_size)
self.o2o = nn.Linear(hidden_size + output_size, output_size)
self.dropout = nn.Dropout(0.1)
self.softmax = nn.LogSoftmax(dim=1)

def forward(self, category, input, hidden):
input_combined = torch.cat((category, input, hidden), 1)
output = self.i2o(input_combined)
hidden = self.i2h(input_combined)
output_combined = torch.cat((output, hidden), 1)
output = self.o2o(output_combined)
output = self.dropout(output)
output = self.softmax(output)
return output, hidden

def initHidden(self):
return Variable(torch.zeros(1, self.hidden_size))

# Training

# Preparing for Training (get random pairs of (category, line))
import random

# Random item from a list
def randomChoice(l):
return l[random.randint(0, len(l) - 1)]

# Get a random category and random line from that category
def randomTrainingPair():
category = randomChoice(all_categories)
line = randomChoice(category_lines[category])
return category, line

# One-hot vector for category
def categoryTensor(category):
li = all_categories.index(category)
tensor = torch.zeros(1, n_categories) # 1*18
tensor[0][li] = 1
return tensor

# One-hot matrix of first to last letters (not including EOS) for input
def inputTensor(line):
tensor = torch.zeros(len(line), 1, n_letters) # len(line)*1*59
for li in range(len(line)):
letter = line[li]
tensor[li][0][all_letters.find(letter)] = 1
return tensor

# LongTensor of second letter to end (EOS) for target
def targetTensor(line):
letter_indexes = [all_letters.find(line[li]) for li in range(1, len(line))]
letter_indexes.append(n_letters - 1) # EOS
return torch.LongTensor(letter_indexes)

# Make category, input, and target tensors from a random category, line pair
def randomTrainingExample():
category, line = randomTrainingPair()
category_tensor = Variable(categoryTensor(category))
input_line_tensor = Variable(inputTensor(line))
target_line_tensor = Variable(targetTensor(line))
return category_tensor, input_line_tensor, target_line_tensor

# our model
rnn = RNN(n_letters, 128, n_letters)

# Training the Network

criterion = nn.NLLLoss()
learning_rate = 0.0005

def train(category_tensor, input_line_tensor, target_line_tensor):
hidden = rnn.initHidden()

rnn.zero_grad()
loss = 0

for i in range(input_line_tensor.size()[0]):
output, hidden = rnn(category_tensor, input_line_tensor[i], hidden)
loss += criterion(output, target_line_tensor[i])

loss.backward()

for p in rnn.parameters():
p.data.add_(-learning_rate, p.grad.data)

return output, loss.data[0] / input_line_tensor.size()[0]

import time
import math

def timeSince(since):
now = time.time()
s = now - since
m = math.floor(s / 60)
s -= m * 60
return '%dm %ds' % (m, s)

n_iters = 100000
print_every = 5000
plot_every = 500
all_losses = []
total_loss = 0 # Reset every plot_every iters

start = time.time()

for iter in range(1, n_iters + 1):
output, loss = train(*randomTrainingExample())
total_loss += loss

if iter % print_every == 0:
print('%s (%d %d%%) %.4f' % (timeSince(start), iter, iter / n_iters * 100, loss))

if iter % plot_every == 0:
all_losses.append(total_loss / plot_every)
total_loss = 0

# 训练loss变化
import matplotlib.pyplot as plt
import matplotlib.ticker as ticker

plt.figure()
plt.plot(all_losses)
plt.show()

# Sample network
max_length = 20
# Sample from a category and starting letter
def sample(category, start_letter='A'):
category_tensor = Variable(categoryTensor(category))
input = Variable(inputTensor(start_letter))
hidden = rnn.initHidden()

output_name = start_letter

# 最大也只生成max_length长度
for i in range(max_length):
output, hidden = rnn(category_tensor, input[0], hidden)
topv, topi = output.data.topk(1)
topi = topi[0][0]
# 如果是EOS，停止
if topi == n_letters - 1:
break
else:
letter = all_letters[topi]
output_name += letter
# 否则，将这个时刻输出的字母作为下个时刻的输入字母
input = Variable(inputTensor(letter))

return output_name

# Get multiple samples from one category and multiple starting letters
def samples(category, start_letters='ABC'):
for start_letter in start_letters:
print(sample(category, start_letter))

samples('Russian', 'RUS')
samples('German', 'GER')
samples('Spanish', 'SPA')
samples('Chinese', 'CHI')

输出结果：
0m 32s (5000 5%) 2.3348
1m 1s (10000 10%) 3.0012
1m 29s (15000 15%) 2.7776
2m 2s (20000 20%) 2.7482
2m 38s (25000 25%) 1.3141
3m 10s (30000 30%) 2.5318
3m 37s (35000 35%) 2.4345
4m 3s (40000 40%) 2.7806
4m 30s (45000 45%) 2.0744
4m 57s (50000 50%) 2.7273
5m 23s (55000 55%) 5.1529
5m 49s (60000 60%) 2.0862
6m 15s (65000 65%) 2.5506
6m 41s (70000 70%) 3.4072
7m 8s (75000 75%) 2.6554
7m 34s (80000 80%) 2.1122
8m 12s (85000 85%) 2.2132
8m 44s (90000 90%) 1.9226
9m 11s (95000 95%) 2.8443
9m 37s (100000 100%) 2.4129
Roskin
Uakinov
Santovov
Gerran
Eren
Romer
Sara
Parez
Aller
Chan
Han
Iun

Process finished with exit code 0

loss变化：



参考：http://pytorch.org/tutorials/intermediate/char_rnn_generation_tutorial.html