基于PyTorch的Seq2Seq翻译模型详细注释介绍(一)
版权声明:本文为博主原创文章,遵循 CC 4.0 BY-SA 版权协议,转载请附上原文出处链接和本声明。
本文链接:https://blog.csdn.net/qysh123/article/details/91245246
Seq2Seq是目前主流的深度学习翻译模型,在自然语言翻译,甚至跨模态知识映射方面都有不错的效果。在软件工程方面,近年来也得到了广泛的应用,例如:
Jiang, Siyuan, Ameer Armaly, and Collin McMillan. “Automatically generating commit messages from diffs using neural machine translation.” In Proceedings of the 32nd IEEE/ACM International Conference on Automated Software Engineering, pp. 135-146. IEEE Press, 2017.
Hu, Xing, Ge Li, Xin Xia, David Lo, and Zhi Jin. “Deep code comment generation.” In Proceedings of the 26th Conference on Program Comprehension, pp. 200-210. ACM, 2018.
这里我结合PyTorch给出的Seq2Seq的示例代码来简单总结一下这个模型实现时的细节以及PyTorch对应的API。PyTorch在其官网上有Tutorial:https://pytorch.org/tutorials/intermediate/seq2seq_translation_tutorial.html,其对应的GitHub链接是:https://github.com/pytorch/tutorials/blob/master/intermediate_source/seq2seq_translation_tutorial.py。这里就以这段代码为例来进行总结:
在上面那个官网的链接中给出了对应数据的下载链接:https://download.pytorch.org/tutorial/data.zip,另外,其实网上很多教程也都是翻译上面这个官方教程的,我也参考了一些,主要包括:
https://www.cnblogs.com/HolyShine/p/9850822.html
https://www.cnblogs.com/www-caiyin-com/p/10123346.html
http://www.pianshen.com/article/5376154542/
所以大家可以以这些教程为基础,我也只是在它们的基础上进行一些补充和解释,所以并不会像上面教程一样给出完整的解释,只是总结一些我觉得重要的内容。首先,初始化编码这些就不总结了,大家看看现有的教程就理解。从Encoder开始总结:
class EncoderRNN(nn.Module):
def __init__(self, input_size, hidden_size):
super(EncoderRNN, self).__init__()#对继承自父类的属性进行初始化。
self.hidden_size = hidden_size
self.embedding = nn.Embedding(input_size, hidden_size)#对输入做初始化Embedding。
self.gru = nn.GRU(hidden_size, hidden_size)#Applies a multilayer gated recurrent unit (GRU) RNN to an input sequence.
def forward(self, input, hidden):
embedded = self.embedding(input).view(1, 1, -1)#view实际上是对现有tensor改造的方法。
output = embedded
output, hidden = self.gru(output, hidden)
return output, hidden
def initHidden(self):
return torch.zeros(1, 1, self.hidden_size, device=device)#初始化,生成(1,1,256)维的全零Tensor。
虽然只有短短几行,可还是有些需要讨论的内容:nn.Embedding是进行初始embedding,当然,这种embedding是完全随机的,并不通过训练或具有实际意义,我觉得网上有些文章连这一点都没搞清楚(例如这里的解释就是错误的:https://my.oschina.net/earnp/blog/1113896),具体可以参看这里的讨论:https://blog.csdn.net/qq_36097393/article/details/88567942。其参数含义可以参考这个解释:nn.Embedding(2, 5),这里的2表示有2个词,5表示维度为5,其实也就是一个2×5的矩阵,所以如果你有1000个词,每个词希望是100维,你就可以这样建立一个word embedding,nn.Embedding(1000, 100)。也可以运行下面我总结示例代码:
import torch
import torch.nn as nn
word_to_ix={\’hello\’:0, \’world\’:1}
embeds=nn.Embedding(2,5)
hello_idx=torch.LongTensor([word_to_ix[\’hello\’]])
world_idx=torch.LongTensor([word_to_ix[\’world\’]])
hello_embed=embeds(hello_idx)
print(hello_embed)
world_embed=embeds(world_idx)
print(world_embed)
具体含义相信大家一看便知,可以试着跑一下(每次print的结果不相同,并且也没啥实际含义)。
另外就是.view(1, 1, -1)的含义,说实话我也没搞清楚过,其实在stackoverflow上已经有人讨论了这个问题:
https://stackoverflow.com/questions/42479902/how-does-the-view-method-work-in-pytorch
大家看看就知,我这里也把上面别人给出的例子提供一下:
import torch
a = torch.range(1, 16)
print(a)
a = a.view(4, 4)
print(a)
Encoder就简单总结这些。下面直接进入到带注意力机制的解码器的总结(为了帮助理解,下面增加了一些注释,说明每一步Tensor的纬度,我个人觉得还是能够便于理解的):
class AttnDecoderRNN(nn.Module):
def __init__(self, hidden_size, output_size, dropout_p=0.1, max_length=MAX_LENGTH):#MAX_LENGTH在翻译任务中定义为10
super(AttnDecoderRNN, self).__init__()
self.hidden_size = hidden_size
self.output_size = output_size#这里的output_size是output_lang.n_words
self.dropout_p = dropout_p#dropout的比例。
self.max_length = max_length
self.embedding = nn.Embedding(self.output_size, self.hidden_size)
self.attn = nn.Linear(self.hidden_size * 2, self.max_length)
self.attn_combine = nn.Linear(self.hidden_size * 2, self.hidden_size)#按照维度要求,进行线性变换。
self.dropout = nn.Dropout(self.dropout_p)
self.gru = nn.GRU(self.hidden_size, self.hidden_size)
self.out = nn.Linear(self.hidden_size, self.output_size)
def forward(self, input, hidden, encoder_outputs):
print(input)
print(\’size of input: \’+str(input.size()))
print(\’size of self.embedding(input): \’+str(self.embedding(input).size()))
embedded = self.embedding(input).view(1, 1, -1)
print(\’size of embedded: \’+str(embedded.size()))
embedded = self.dropout(embedded)
print(\’size of embedded[0]: \’+str(embedded[0].size()))
print(\’size of torch.cat((embedded[0], hidden[0]), 1): \’+str(torch.cat((embedded[0], hidden[0]), 1).size()))
print(\’size of self.attn(torch.cat((embedded[0], hidden[0]), 1)): \’+str(self.attn(torch.cat((embedded[0], hidden[0]), 1)).size()))
#Size of embedded: [1,1,256]
#Size of embedded[0]: [1,256]
#Size of size of torch.cat((embedded[0], hidden[0]), 1): [1,512]
# 此处相当于学出来了attention的权重
# 需要注意的是torch的concatenate函数是torch.cat,是在已有的维度上拼接,按照代码中的写法,就是在第二个纬度上拼接。
# 而stack是建立一个新的维度,然后再在该纬度上进行拼接。
attn_weights = F.softmax(
self.attn(torch.cat((embedded[0], hidden[0]), 1)), dim=1)#这里的F.softmax表示的是torch.nn.functional.softmax
#Size of attn_weights: [1,10]
#Size of attn_weights.unsqueeze(0): [1,1,10]
#Size of encoder_outputs: [10,256]
#Size of encoder_outputs.unsqueeze(0): [1,10,256]
#unsqueeze的解释是Returns a new tensor with a dimension of size one inserted at the specified position.
attn_applied = torch.bmm(attn_weights.unsqueeze(0),
encoder_outputs.unsqueeze(0))#bmm本质上来讲是个批量的矩阵乘操作。
#Size of attn_applied: [1,1,256]
output = torch.cat((embedded[0], attn_applied[0]), 1)
#Size of output here is: [1,512]
print(\’size of output (at this location): \’+str(output.size()))
output = self.attn_combine(output).unsqueeze(0)
#Size of output here is: [1,1,256]
#print(output)
output = F.relu(output)#rectified linear unit function element-wise:
#print(output)
output, hidden = self.gru(output, hidden)
output = F.log_softmax(self.out(output[0]), dim=1)
print(\’\’)
print(\’————\’)
return output, hidden, attn_weights
def initHidden(self):
return torch.zeros(1, 1, self.hidden_size, device=device)
首先是dropout,关于dropout可以首先参考一下PyTorch的官方解释:
https://pytorch.org/docs/stable/nn.html?highlight=nn%20dropout#torch.nn.Dropout
简单来说,就是During training, randomly zeroes some of the elements of the input tensor with probability p using samples from a Bernoulli distribution,有朋友给出了很详细的讨论和解释:
https://blog.csdn.net/stdcoutzyx/article/details/49022443
其次应该注意一下nn.Linear的含义和作用,还是给出官网的解释:Applies a linear transformation to the incoming data,类似地,可以参考一下我下面给出的示例代码:
import torch
import torch.nn as nn
m = nn.Linear(2, 3)
input = torch.randn(2, 2)
print(input)
output = m(input)
print(output)
接下来解释一下torch.bmm。按照PyTorch官网的解释,https://pytorch.org/docs/stable/torch.html?highlight=torch%20bmm#torch.bmm
torch.bmm起的作用是:Performs a batch matrix-matrix product of matrices stored in batch1 and batch2,这样的解释还是太抽象,其实通过一个例子就很好懂了,实际就是一个批量矩阵乘法:
import torch
batch1=torch.randn(2,3,4)
print(batch1)
batch2=torch.randn(2,4,5)
print(batch2)
res=torch.bmm(batch1,batch2)
print(res)
具体的乘法规则是:If batch1 is a (b×n×m) tensor, batch2 is a (b×m×p) tensor, out will be a (b×n×p) tensor.
关于torch.cat,还是以PyTorch官网给出的例子做一个简单说明:
Concatenates the given sequence of seq tensors in the given dimension. 例子如下:
import torch
x=torch.randn(2,3)
print(x)
print(torch.cat((x, x, x), 0))
print(torch.cat((x, x, x), 1))
这里就先总结到这里,会在下一篇博客中继续总结。
————————————————
版权声明:本文为CSDN博主「蛐蛐蛐」的原创文章,遵循 CC 4.0 BY-SA 版权协议,转载请附上原文出处链接及本声明。
原文链接:https://blog.csdn.net/qysh123/article/details/91245246