keras理解词嵌入层 [英] keras understanding Word Embedding Layer
问题描述
在页面中,我得到了以下代码:
From the page I got the below code:
from numpy import array
from keras.preprocessing.text import one_hot
from keras.preprocessing.sequence import pad_sequences
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import Flatten
from keras.layers.embeddings import Embedding
# define documents
docs = ['Well done!',
'Good work',
'Great effort',
'nice work',
'Excellent!',
'Weak',
'Poor effort!',
'not good',
'poor work',
'Could have done better.']
# define class labels
labels = array([1,1,1,1,1,0,0,0,0,0])
# integer encode the documents
vocab_size = 50
encoded_docs = [one_hot(d, vocab_size) for d in docs]
print(encoded_docs)
# pad documents to a max length of 4 words
max_length = 4
padded_docs = pad_sequences(encoded_docs, maxlen=max_length, padding='post')
print(padded_docs)
# define the model
model = Sequential()
model.add(Embedding(vocab_size, 8, input_length=max_length))
model.add(Flatten())
model.add(Dense(1, activation='sigmoid'))
# compile the model
model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['acc'])
# summarize the model
print(model.summary())
# fit the model
model.fit(padded_docs, labels, epochs=50, verbose=0)
# evaluate the model
loss, accuracy = model.evaluate(padded_docs, labels, verbose=0)
print('Accuracy: %f' % (accuracy*100))
- 我查看了
encoded_docs
,发现单词done
和work
都具有one_hot编码为2的代码,为什么?是因为unicity of word to index mapping non-guaranteed.
与此页面一致? - 我通过命令
embeddings = model.layers[0].get_weights()[0]
获得了embeddings
.在这种情况下,为什么我们得到大小为50的embedding
对象?即使两个单词具有相同的one_hot数字,但它们的嵌入方式是否不同? - 我怎么能理解哪个单词是哪个嵌入,即
done
vswork
-
我还在页面上找到了以下代码可能有助于查找每个单词的嵌入.但是我不知道如何创建
word_to_index
- I looked at
encoded_docs
and noticed that wordsdone
andwork
both have one_hot encoding of 2, why? Is it becauseunicity of word to index mapping non-guaranteed.
as per this page? - I got
embeddings
by commandembeddings = model.layers[0].get_weights()[0]
. in such case why do we getembedding
object of size 50? Even though two words have same one_hot number, do they have different embedding? - how could i understand which embedding is for which word i.e.
done
vswork
I also found below code at the page that could help with finding embedding of each word. But i dont know how to create
word_to_index
word_to_index
是从单词到其索引的映射(即dict). love
:69
words_embeddings = {w:embeddings [idx] for w,word_to_index.items()中的idx}
word_to_index
is a mapping (i.e. dict) from words to their index, e.g. love
: 69
words_embeddings = {w:embeddings[idx] for w, idx in word_to_index.items()}
请确保我对para #
的理解是正确的.
Please ensure that my understanding of para #
is correct.
第一层具有400个参数,因为总字数为50,并且嵌入具有8个维度,因此50 * 8 = 400.
The first layer has 400 parameters because total word count is 50 and embedding have 8 dimensions so 50*8=400.
最后一层具有33个参数,因为每个句子最多包含4个单词.因此,由于嵌入尺寸为4 * 8,偏置为1.共33个
The last layer has 33 parameters because each sentence has 4 words max. So 4*8 due to dimensions of embedding and 1 for bias. 33 total
_________________________________________________________________
Layer (type) Output Shape Param#
=================================================================
embedding_3 (Embedding) (None, 4, 8) 400
_________________________________________________________________
flatten_3 (Flatten) (None, 32) 0
_________________________________________________________________
dense_3 (Dense) (None, 1) 33
=================================================================
- 最后,如果上面的1是正确的,是否有一种更好的方法来获得嵌入层
model.add(Embedding(vocab_size, 8, input_length=max_length))
而无需进行一次热编码encoded_docs = [one_hot(d, vocab_size) for d in docs]
- Finally, if 1 above is correct, is there a better way to get embedding layer
model.add(Embedding(vocab_size, 8, input_length=max_length))
without doing one hot codingencoded_docs = [one_hot(d, vocab_size) for d in docs]
++++++++++++++++++++++++++++++++++ 更新-提供更新的代码
+++++++++++++++++++++++++++++++ update - providing the updated code
from numpy import array
from keras.preprocessing.text import one_hot
from keras.preprocessing.sequence import pad_sequences
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import Flatten
from keras.layers.embeddings import Embedding
# define documents
docs = ['Well done!',
'Good work',
'Great effort',
'nice work',
'Excellent!',
'Weak',
'Poor effort!',
'not good',
'poor work',
'Could have done better.']
# define class labels
labels = array([1,1,1,1,1,0,0,0,0,0])
from keras.preprocessing.text import Tokenizer
tokenizer = Tokenizer()
#this creates the dictionary
#IMPORTANT: MUST HAVE ALL DATA - including Test data
#IMPORTANT2: This method should be called only once!!!
tokenizer.fit_on_texts(docs)
#this transforms the texts in to sequences of indices
encoded_docs2 = tokenizer.texts_to_sequences(docs)
encoded_docs2
max_length = 4
padded_docs2 = pad_sequences(encoded_docs2, maxlen=max_length, padding='post')
max_index = array(padded_docs2).reshape((-1,)).max()
# define the model
model = Sequential()
model.add(Embedding(max_index+1, 8, input_length=max_length))# you cannot use just max_index
model.add(Flatten())
model.add(Dense(1, activation='sigmoid'))
# compile the model
model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['acc'])
# summarize the model
print(model.summary())
# fit the model
model.fit(padded_docs2, labels, epochs=50, verbose=0)
# evaluate the model
loss, accuracy = model.evaluate(padded_docs2, labels, verbose=0)
print('Accuracy: %f' % (accuracy*100))
embeddings = model.layers[0].get_weights()[0]
embeding_for_word_7 = embeddings[14]
index = tokenizer.texts_to_sequences([['well']])[0][0]
tokenizer.document_count
tokenizer.word_index
推荐答案
1 - Yes, word unicity is not guaranteed, see the docs:
- 来自
one_hot
:这是hashing_trick
函数的包装... - 来自
hashing_trick
:由于散列函数可能发生的冲突,可能会将两个或多个单词分配给同一索引.散列空间和不同对象的数量."
- From
one_hot
: This is a wrapper to thehashing_trick
function... - From
hashing_trick
: "Two or more words may be assigned to the same index, due to possible collisions by the hashing function. The probability of a collision is in relation to the dimension of the hashing space and the number of distinct objects."
为此最好使用Tokenizer
. (请参阅问题4)
It would be better to use a Tokenizer
for this. (See question 4)
切记非常重要,在创建索引时应同时包含所有所有单词.您不能使用函数来创建包含2个单词的词典,然后再创建2个单词,再创建一个词典.这将创建非常错误的词典.
It's very important to remember that you should involve all words at once when creating indices. You cannot use a function to create a dictionary with 2 words, then again with 2 words, then again.... This will create very wrong dictionaries.
2-嵌入的大小为50 x 8
,因为它是在嵌入层中定义的:
2 - Embeddings have the size 50 x 8
, because that was defined in the embedding layer:
Embedding(vocab_size, 8, input_length=max_length)
-
vocab_size = 50
-这意味着词典中有50个单词 -
embedding_size= 8
-这是嵌入的真实大小:每个单词由8个数字的向量表示. vocab_size = 50
- this means there are 50 words in the dictionaryembedding_size= 8
- this is the true size of the embedding: each word is represented by a vector of 8 numbers.
3-您不知道.它们使用相同的嵌入.
3 - You don't know. They use the same embedding.
系统将使用相同的嵌入(索引为2的嵌入).这对于您的模型根本不健康.您应该使用另一种方法来创建问题1中的索引.
The system will use the same embedding (the one for index = 2). This is not healthy for your model at all. You should use another method for creating indices in question 1.
4-您可以手动创建单词词典,也可以使用Tokenizer
类.
4 - You can create a word dictionary manually, or use the Tokenizer
class.
手动:
确保删除标点符号,并将所有单词都小写.
Make sure you remove punctuation, make all words lower case.
只需为您拥有的每个单词创建字典:
Just create a dictionary for each word you have:
dictionary = dict()
current_key = 1
for doc in docs:
for word in doc.split(' '):
#make sure you remove punctuation (this might be boring)
word = word.lower()
if not (word in dictionary):
dictionary[word] = current_key
current_key += 1
令牌生成器:
from keras.preprocessing.text import Tokenizer
tokenizer = Tokenizer()
#this creates the dictionary
#IMPORTANT: MUST HAVE ALL DATA - including Test data
#IMPORTANT2: This method should be called only once!!!
tokenizer.fit_on_texts(docs)
#this transforms the texts in to sequences of indices
encoded_docs2 = tokenizer.texts_to_sequences(docs)
请参见encoded_docs2
的输出:
[[6, 2], [3, 1], [7, 4], [8, 1], [9], [10], [5, 4], [11, 3], [5, 1], [12, 13, 2, 14]]
查看最大索引:
padded_docs2 = pad_sequences(encoded_docs2, maxlen=max_length, padding='post')
max_index = array(padded_docs2).reshape((-1,)).max()
因此,您的vocab_size
应该为15(否则,您将有很多无用且无害的嵌入行).请注意,0
未被用作索引.它将显示在填充中!!!
So, your vocab_size
should be 15 (otherwise you'd have lots of useless - and harmless - embedding rows). Notice that 0
was not used as an index. It will appear in padding!!!
不要再次适合"令牌生成器!仅使用
texts_to_sequences()
或其他方法此处与拟合"无关.
Do not "fit" the tokenizer again! Only use
texts_to_sequences()
or other methods here that are not related to "fitting".
提示::有时在文本中包含end_of_sentence
个单词可能会很有用.
Hint: it might be useful to include end_of_sentence
words in your text sometimes.
提示2::保存Tokenizer
以便以后使用是一个好主意(因为它具有用fit_on_texts
创建的特定数据字典).
Hint2: it is a good idea to save your Tokenizer
to be used later (since it has a specific dictoinary for your data, created with fit_on_texts
).
#save:
text_to_save = tokenizer.to_json()
#load:
from keras.preprocessing.text import tokenizer_from_json
tokenizer = tokenizer_from_json(loaded_text)
5-嵌入参数正确.
5 - Params for embedding are correct.
密度:
Dense
的参数始终基于上一层(在本例中为Flatten
).
Params for Dense
are always based on the preceding layer (the Flatten
in this case).
公式为:previous_output * units + units
这将导致32 (from the Flatten) * 1 (Dense units) + 1 (Dense bias=units) = 33
展平:
将所有先前的尺寸乘以= 8 * 4
.
Embedding
输出lenght = 4
和embedding_size = 8
.
It gets all the previous dimensions multiplied = 8 * 4
.
The Embedding
outputs lenght = 4
and embedding_size = 8
.
6-Embedding
层与数据以及预处理方式无关.
6 - The Embedding
layer is not dependent of your data and how you preprocess it.
Embedding
层的大小仅为50 x 8,因为您已经这样说过. (请参阅问题2)
The Embedding
layer has simply the size 50 x 8 because you told so. (See question 2)
当然,还有更好的预处理数据的方法-请参阅问题4.
There are, of course, better ways of preprocessing the data - See question 4.
这将使您选择更好的vocab_size
(即字典大小).
This will lead you to select better the vocab_size
(which is dictionary size).
获取嵌入矩阵:
embeddings = model.layers[0].get_weights()[0]
选择任何单词索引:
embeding_for_word_7 = embeddings[7]
仅此而已.
如果您使用的是分词器,请使用以下命令获取单词index:
If you're using a tokenizer, get the word index with:
index = tokenizer.texts_to_sequences([['word']])[0][0]
这篇关于keras理解词嵌入层的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!