利用TfidfVectorizer進行中文文字分類（資料集是復旦中文語料）
和之前介紹的不同，重構了些程式碼，為了使整個流程更加清楚，我們要重新對資料進行預處理。
閱讀本文，你可以瞭解中文文字分類從資料預處理、模型定義、訓練和測試的整個流程。
一、熟悉資料
資料的格式是這樣子的：
基本目錄如下：
其中train存放的是訓練集，answer存放的是測試集，具體看下train中的檔案：
下面有20個資料夾，對應著20個類，我們繼續看下其中的檔案，以C3-Art為例：
每一篇都對應著一個txt檔案，編碼格式是gb18030.utf8資料夾下的是utf-8編碼格式的txt檔案。
其中C3-Art0001.txt的部分內容如下：
二、資料預處理
本文資料預處理基本流程：
- 先將所有訓練資料的txt路徑以及測試用的txt路徑寫入到txt中備用：train.txt、test.txt
def _txtpath_to_txt(self): #將訓練集和測試集下的txt路徑儲存 train_txt_path = os.path.join(PATH, "process/Fudan/train.txt") test_txt_path = os.path.join(PATH, "process/Fudan//test.txt") train_list = os.listdir(os.path.join(PATH, self.trainPath)) #獲得該目錄下的所有資料夾，返回一個列表 fp1 = open(train_txt_path,"w",encoding="utf-8") fp2 = open(test_txt_path,"w",encoding="utf-8") for train_dir in train_list: #取得下一級目錄下的所有的txt路徑（絕對路徑） for txt in glob.glob(os.path.join(PATH,self.trainPath+train_dir+"/*.txt")): fp1.write(txt+"\n") fp1.close() test_list = os.listdir(os.path.join(PATH,self.testPath)) #獲得該目錄下的所有資料夾，返回一個列表 for test_dir in test_list: for txt in glob.glob(os.path.join(PATH, self.testPath+test_dir+"/*.txt")): fp2.write(txt+"\n") fp2.close()
- 接下來我們要將txt中的文字分詞後寫入到分詞文字txt中，以及對應的標籤寫入到標籤txt中：train_content.txt、train_label.txt、test_content.txt、test_label.txt
#將txt中的文字和標籤儲存到txt中 def _contentlabel_to_txt(self, txt_path, content_path, label_path): files = open(txt_path,"r",encoding="utf-8") content_file = open(content_path,"w",encoding="utf-8") label_file = open(label_path,"w",encoding="utf-8") for txt in files.readlines(): #讀取每一行的txt txt = txt.strip() #去除掉\n content_list=[] label_str = txt.split("/")[-1].split("-")[-1] #先用/進行切割，獲取列表中的最後一個，再利用-進行切割，獲取最後一個 label_list = [] #以下for迴圈用於獲取標籤，遍歷每個字元，如果遇到了數字，就終止 for s in label_str: if s.isalpha(): label_list.append(s) elif s.isalnum(): break else: print("出錯了") label = "".join(label_list) #將字元列表轉換為字串，得到標籤 #print(label) #以下用於獲取所有文字 fp1 = open(txt,"r",encoding="gb18030",errors='ignore') #以gb18030的格式開啟檔案，errors='ignore'用於忽略掉超過該字元編碼範圍的字元 for line in fp1.readlines(): #讀取每一行 #jieba分詞，精確模式 line = jieba.lcut(line.strip(), cut_all=False) #將每一行分詞的結果儲存在一個list中 content_list.extend(line) fp1.close() content_str = " ".join(content_list) #轉成字串 #print(content_str) content_file.write(content_str+"\n") #將文字儲存到tx中 label_file.write(label+"\n") content_file.close() label_file.close() files.close()
儲存的分詞後的文字是這個樣子的：
標籤是這樣子的：
- 接下來我們將訓練用的分詞文字和測試用的分詞文字進行合併後利用word2vec訓練詞向量（不用過濾掉停止詞）：
from gensim.models import Word2Vec from gensim.models.word2vec import PathLineSentences import multiprocessing import os import sys import logging # 日誌資訊輸出 program = os.path.basename(sys.argv[0]) logger = logging.getLogger(program) logging.basicConfig(format='%(asctime)s: %(levelname)s: %(message)s') logging.root.setLevel(level=logging.INFO) logger.info("running %s" % ' '.join(sys.argv)) # check and process input arguments # if len(sys.argv) < 4: # print(globals()['__doc__'] % locals()) # sys.exit(1) # input_dir, outp1, outp2 = sys.argv[1:4] # 訓練模型 # 輸入語料目錄:PathLineSentences(input_dir) # embedding size:200 共現視窗大小:10 去除出現次數10以下的詞,多執行緒執行,迭代10次 model = Word2Vec(PathLineSentences('/content/drive/My Drive/transformer/process/Fudan/word2vec/data/'), size=200, window=10, min_count=10, workers=multiprocessing.cpu_count(), iter=10) model.save('/content/drive/My Drive/transformer/process/Fudan/word2vec/model/Word2vec.w2v')
'/content/drive/My Drive/transformer/Fudan/word2vec/data/'下是train_content.txt和test_content.txt（將它們移過去的）
最後會生成：
- 去掉停用詞
#去除掉停用詞 def _get_clean_data(self, filePath): #先初始化停用詞字典 self._get_stopwords() sentence_list = [] with open(filePath,'r',encoding='utf-8') as fp: lines = fp.readlines() for line in lines: tmp = [] words = line.strip().split(" ") for word in words: word = word.strip() if word not in self.stopWordDict and word != '': tmp.append(word) else: continue sentence_list.append(tmp) return sentence_list #讀取停用詞字典 def _get_stopwords(self): with open(os.path.join(PATH, self.stopWordSource), "r") as f: stopWords = f.read() stopWordList = set(stopWords.splitlines()) # 將停用詞用列表的形式生成，之後查詢停用詞時會比較快 self.stopWordDict = dict(zip(stopWordList, list(range(len(stopWordList)))))
- 建立詞彙表
#建立詞彙表 def _get_vocaburay(self): train_content = os.path.join(PATH, "process/Fudan/word2vec/data/train_content.txt") sentence_list = self._get_clean_data(train_content) #這裡可以計算文字的平均長度，設定配置中的sequenceLength #max_sequence = sum([len(s) for s in sentence_list]) / len(sentence_list) vocab_before = [] for sentence in sentence_list: for word in sentence: vocab_before.append(word) count_vocab = Counter(vocab_before) #統計每個詞出現的次數 #print(len(count_vocab)) count_vocab = sorted(count_vocab.items(),key=lambda x:x[1], reverse=True) #將出現頻率按從高到低排序 vocab_after = copy.deepcopy(count_vocab[:6000]) return dict(vocab_after) #返回前6000個詞，將元組構成的列表轉換為字典
-
將文字轉換為id，將標籤轉換為Id
def _wordToIdx(self): #構建詞彙和id的對映 vocab = list(self._get_vocaburay().keys()) #取得字典中的鍵，也就是詞語，轉換成列表 #print(vocab) tmp = ['PAD','UNK'] vocab = tmp + vocab word2idx = {word:i for i,word in enumerate(vocab)} idx2word = {i:word for i,word in enumerate(vocab)} return word2idx,idx2word def _labelToIdx(self): #構建詞彙列表和到id的對映 label_path = os.path.join(PATH, "process/Fudan/train_label.txt") with open(os.path.join(PATH, label_path), "r") as f: labels = f.read() labelsList = sorted(set(labels.splitlines())) #為了避免每次標籤id變換，這裡排個序 label2idx = {label:i for i,label in enumerate(labelsList)} idx2label = {i:label for i,label in enumerate(labelsList)} self.labelList = [label2idx[label] for label in labelsList] return label2idx,idx2label
- 獲取訓練資料
def _getData(self,contentPath,labelPath,mode=None): #這裡有兩種操作，如果文字中的詞沒有在詞彙表中出現，則可以捨去或者用UNK代替，我們這裡使用UNK vocab = self._get_vocaburay() word2idx,idx2word = self._wordToIdx() label2idx,idx2label = self._labelToIdx() data = [] content_list = self._get_clean_data(contentPath) for content in content_list: #print(content) tmp = [] if len(content) >= self.config.sequenceLength: #大於最大長度進行截斷 content = content[:self.config.sequenceLength] else: #小於最大長度用PAD的id進行填充層 content = ['PAD']*(self.config.sequenceLength-len(content)) + content for word in content: #將詞語用id進行對映 if word in word2idx: tmp.append(word2idx[word]) else: tmp.append(word2idx['UNK']) data.append(tmp) with open(labelPath,'r',encoding='utf-8') as fp: labels = fp.read() label = [[label2idx[label]] for label in labels.splitlines()] return data,label
- 將訓練資料拆分為訓練集和驗證集
def _getData(self,contentPath,labelPath,mode=None): #這裡有兩種操作，如果文字中的詞沒有在詞彙表中出現，則可以捨去或者用UNK代替，我們這裡使用UNK vocab = self._get_vocaburay() word2idx,idx2word = self._wordToIdx() label2idx,idx2label = self._labelToIdx() data = [] content_list = self._get_clean_data(contentPath) for content in content_list: #print(content) tmp = [] if len(content) >= self.config.sequenceLength: #大於最大長度進行截斷 content = content[:self.config.sequenceLength] else: #小於最大長度用PAD的id進行填充層 content = ['PAD']*(self.config.sequenceLength-len(content)) + content for word in content: #將詞語用id進行對映 if word in word2idx: tmp.append(word2idx[word]) else: tmp.append(word2idx['UNK']) data.append(tmp) with open(labelPath,'r',encoding='utf-8') as fp: labels = fp.read() label = [[label2idx[label]] for label in labels.splitlines()] return data,label
- 獲取測試集
def _getTrainValData(self,dataPath,labelPath): trainData,trainLabel = self._getData(dataPath,labelPath) #方便起見，我們這裡就直接使用sklearn中的函式了 self.trainData,self.valData,self.trainLabels,self.valLabels = train_test_split(trainData,trainLabel,test_size=self.rate,random_state=1) def _getTestData(self,dataPath,labelPath): self.testData,self.testLabels = self._getData(dataPath,labelPath)
- 獲得詞向量
#獲取詞彙表中的詞向量 def _getWordEmbedding(self): word2idx,idx2word = self._wordToIdx() vocab = sorted(word2idx.items(), key=lambda x:x[1]) #將詞按照id進行排序 #print(vocab) w2vModel = Word2Vec.load(os.path.join(PATH,'process/Fudan/word2vec/model/Word2vec.w2v')) self.wordEmbedding.append([0]*self.embeddingSize) #PAD對應的詞向量 self.wordEmbedding.append([0]*self.embeddingSize) #UNK對應的詞向量 for i in range(2,len(vocab)): self.wordEmbedding.append(list(w2vModel[vocab[i][0]]))
程式碼就不一一貼了。
三、建立模型
import numpy as np import tensorflow as tf import warnings warnings.filterwarnings("ignore") class Transformer(object): """ Transformer Encoder 用於文字分類 """ def __init__(self, config, wordEmbedding): # 定義模型的輸入 #inputX:[None,600]，inputY:[None,20] self.inputX = tf.placeholder(tf.int32, [None, config.sequenceLength], name="inputX") self.inputY = tf.placeholder(tf.int32, [None, config.numClasses], name="inputY") self.lastBatch = False self.dropoutKeepProb = tf.placeholder(tf.float32, name="dropoutKeepProb") self.config = config # 定義l2損失 l2Loss = tf.constant(0.0) # 詞嵌入層, 位置向量的定義方式有兩種：一是直接用固定的one-hot的形式傳入，然後和詞向量拼接，在當前的資料集上表現效果更好。另一種 # 就是按照論文中的方法實現，這樣的效果反而更差，可能是增大了模型的複雜度，在小資料集上表現不佳。 with tf.name_scope("wordEmbedding"): self.W = tf.Variable(tf.cast(wordEmbedding, dtype=tf.float32, name="word2vec"), name="W") self.wordEmbedded = tf.nn.embedding_lookup(self.W, self.inputX) with tf.name_scope("positionEmbedding"): if tf.shape(self.wordEmbedded)[0] == config.batchSize: self.positionEmbedded = self._positionEmbedding() else: self.positionEmbedded = self._positionEmbedding(lastBatch=tf.shape(self.wordEmbedded)[0]) self.embeddedWords = self.wordEmbedded + self.positionEmbedded with tf.name_scope("transformer"): for i in range(config.modelConfig.numBlocks): with tf.name_scope("transformer-{}".format(i + 1)): # 維度[batch_size, sequence_length, embedding_size] multiHeadAtt = self._multiheadAttention(rawKeys=self.wordEmbedded, queries=self.embeddedWords, keys=self.embeddedWords) # 維度[batch_size, sequence_length, embedding_size] self.embeddedWords = self._feedForward(multiHeadAtt, [config.modelConfig.filters, config.modelConfig.embeddingSize]) outputs = tf.reshape(self.embeddedWords, [-1, config.sequenceLength * (config.modelConfig.embeddingSize)]) outputSize = outputs.get_shape()[-1].value with tf.name_scope("dropout"): outputs = tf.nn.dropout(outputs, keep_prob=self.dropoutKeepProb) # 全連線層的輸出 with tf.name_scope("output"): outputW = tf.get_variable( "outputW", shape=[outputSize, config.numClasses], initializer=tf.contrib.layers.xavier_initializer()) outputB= tf.Variable(tf.constant(0.1, shape=[config.numClasses]), name="outputB") l2Loss += tf.nn.l2_loss(outputW) l2Loss += tf.nn.l2_loss(outputB) self.logits = tf.nn.xw_plus_b(outputs, outputW, outputB, name="logits") if config.numClasses == 1: self.predictions = tf.cast(tf.greater_equal(self.logits, 0.0), tf.float32, name="predictions") elif config.numClasses > 1: self.predictions = tf.argmax(self.logits, axis=-1, name="predictions") # 計算二元交叉熵損失 with tf.name_scope("loss"): if config.numClasses == 1: losses = tf.nn.sigmoid_cross_entropy_with_logits(logits=self.logits, labels=tf.cast(tf.reshape(self.inputY, [-1, 1]), dtype=tf.float32)) elif config.numClasses > 1: print(self.logits,self.inputY) losses = tf.nn.softmax_cross_entropy_with_logits(logits=self.logits, labels=self.inputY) self.loss = tf.reduce_mean(losses) + config.modelConfig.l2RegLambda * l2Loss def _layerNormalization(self, inputs, scope="layerNorm"): # LayerNorm層和BN層有所不同 epsilon = self.config.modelConfig.epsilon inputsShape = inputs.get_shape() # [batch_size, sequence_length, embedding_size] paramsShape = inputsShape[-1:] # LayerNorm是在最後的維度上計算輸入的資料的均值和方差，BN層是考慮所有維度的 # mean, variance的維度都是[batch_size, sequence_len, 1] mean, variance = tf.nn.moments(inputs, [-1], keep_dims=True) beta = tf.Variable(tf.zeros(paramsShape)) gamma = tf.Variable(tf.ones(paramsShape)) normalized = (inputs - mean) / ((variance + epsilon) ** .5) outputs = gamma * normalized + beta return outputs def _multiheadAttention(self, rawKeys, queries, keys, numUnits=None, causality=False, scope="multiheadAttention"): # rawKeys 的作用是為了計算mask時用的，因為keys是加上了position embedding的，其中不存在padding為0的值 numHeads = self.config.modelConfig.numHeads keepProp = self.config.modelConfig.keepProp if numUnits is None: # 若是沒傳入值，直接去輸入資料的最後一維，即embedding size. numUnits = queries.get_shape().as_list()[-1] # tf.layers.dense可以做多維tensor資料的非線性對映，在計算self-Attention時，一定要對這三個值進行非線性對映， # 其實這一步就是論文中Multi-Head Attention中的對分割後的資料進行權重對映的步驟，我們在這裡先對映後分割，原則上是一樣的。 # Q, K, V的維度都是[batch_size, sequence_length, embedding_size] Q = tf.layers.dense(queries, numUnits, activation=tf.nn.relu) K = tf.layers.dense(keys, numUnits, activation=tf.nn.relu) V = tf.layers.dense(keys, numUnits, activation=tf.nn.relu) # 將資料按最後一維分割成num_heads個, 然後按照第一維拼接 # Q, K, V 的維度都是[batch_size * numHeads, sequence_length, embedding_size/numHeads] Q_ = tf.concat(tf.split(Q, numHeads, axis=-1), axis=0) K_ = tf.concat(tf.split(K, numHeads, axis=-1), axis=0) V_ = tf.concat(tf.split(V, numHeads, axis=-1), axis=0) # 計算keys和queries之間的點積，維度[batch_size * numHeads, queries_len, key_len], 後兩維是queries和keys的序列長度 similary = tf.matmul(Q_, tf.transpose(K_, [0, 2, 1])) # 對計算的點積進行縮放處理，除以向量長度的根號值 scaledSimilary = similary / (K_.get_shape().as_list()[-1] ** 0.5) # 在我們輸入的序列中會存在padding這個樣的填充詞，這種詞應該對最終的結果是毫無幫助的，原則上說當padding都是輸入0時， # 計算出來的權重應該也是0，但是在transformer中引入了位置向量，當和位置向量相加之後，其值就不為0了，因此在新增位置向量 # 之前，我們需要將其mask為0。雖然在queries中也存在這樣的填充詞，但原則上模型的結果之和輸入有關，而且在self-Attention中 # queryies = keys，因此只要一方為0，計算出的權重就為0。 # 具體關於key mask的介紹可以看看這裡： https://github.com/Kyubyong/transformer/issues/3 # 利用tf，tile進行張量擴張， 維度[batch_size * numHeads, keys_len] keys_len = keys 的序列長度 # tf.tile((?, 200), [8,1]) # 將每一時序上的向量中的值相加取平均值 keyMasks = tf.sign(tf.abs(tf.reduce_sum(rawKeys, axis=-1))) # 維度[batch_size, time_step] print(keyMasks.shape) keyMasks = tf.tile(keyMasks, [numHeads, 1]) # 增加一個維度，並進行擴張，得到維度[batch_size * numHeads, queries_len, keys_len] keyMasks = tf.tile(tf.expand_dims(keyMasks, 1), [1, tf.shape(queries)[1], 1]) # tf.ones_like生成元素全為1，維度和scaledSimilary相同的tensor, 然後得到負無窮大的值 paddings = tf.ones_like(scaledSimilary) * (-2 ** (32 + 1)) # tf.where(condition, x, y),condition中的元素為bool值，其中對應的True用x中的元素替換，對應的False用y中的元素替換 # 因此condition,x,y的維度是一樣的。下面就是keyMasks中的值為0就用paddings中的值替換 maskedSimilary = tf.where(tf.equal(keyMasks, 0), paddings, scaledSimilary) # 維度[batch_size * numHeads, queries_len, key_len] # 在計算當前的詞時，只考慮上文，不考慮下文，出現在Transformer Decoder中。在文字分類時，可以只用Transformer Encoder。 # Decoder是生成模型，主要用在語言生成中 if causality: diagVals = tf.ones_like(maskedSimilary[0, :, :]) # [queries_len, keys_len] tril = tf.contrib.linalg.LinearOperatorTriL(diagVals).to_dense() # [queries_len, keys_len] masks = tf.tile(tf.expand_dims(tril, 0), [tf.shape(maskedSimilary)[0], 1, 1]) # [batch_size * numHeads, queries_len, keys_len] paddings = tf.ones_like(masks) * (-2 ** (32 + 1)) maskedSimilary = tf.where(tf.equal(masks, 0), paddings, maskedSimilary) # [batch_size * numHeads, queries_len, keys_len] # 通過softmax計算權重係數，維度 [batch_size * numHeads, queries_len, keys_len] weights = tf.nn.softmax(maskedSimilary) # 加權和得到輸出值, 維度[batch_size * numHeads, sequence_length, embedding_size/numHeads] outputs = tf.matmul(weights, V_) # 將多頭Attention計算的得到的輸出重組成最初的維度[batch_size, sequence_length, embedding_size] outputs = tf.concat(tf.split(outputs, numHeads, axis=0), axis=2) outputs = tf.nn.dropout(outputs, keep_prob=keepProp) # 對每個subLayers建立殘差連線，即H(x) = F(x) + x outputs += queries # normalization 層 outputs = self._layerNormalization(outputs) return outputs def _feedForward(self, inputs, filters, scope="multiheadAttention"): # 在這裡的前向傳播採用卷積神經網路 # 內層 params = {"inputs": inputs, "filters": filters[0], "kernel_size": 1, "activation": tf.nn.relu, "use_bias": True} outputs = tf.layers.conv1d(**params) # 外層 params = {"inputs": outputs, "filters": filters[1], "kernel_size": 1, "activation": None, "use_bias": True} # 這裡用到了一維卷積，實際上卷積核尺寸還是二維的，只是只需要指定高度，寬度和embedding size的尺寸一致 # 維度[batch_size, sequence_length, embedding_size] outputs = tf.layers.conv1d(**params) # 殘差連線 outputs += inputs # 歸一化處理 outputs = self._layerNormalization(outputs) return outputs def _positionEmbedding(self, lastBatch=None, scope="positionEmbedding"): # 生成可訓練的位置向量 if lastBatch is None: batchSize = self.config.batchSize #128 else: batchSize = lastBatch sequenceLen = self.config.sequenceLength #600 embeddingSize = self.config.modelConfig.embeddingSize #100 # 生成位置的索引，並擴張到batch中所有的樣本上 positionIndex = tf.tile(tf.expand_dims(tf.range(sequenceLen), 0), [batchSize, 1]) # 根據正弦和餘弦函式來獲得每個位置上的embedding的第一部分 positionEmbedding = np.array([[pos / np.power(10000, (i-i%2) / embeddingSize) for i in range(embeddingSize)] for pos in range(sequenceLen)]) # 然後根據奇偶性分別用sin和cos函式來包裝 positionEmbedding[:, 0::2] = np.sin(positionEmbedding[:, 0::2]) positionEmbedding[:, 1::2] = np.cos(positionEmbedding[:, 1::2]) # 將positionEmbedding轉換成tensor的格式 positionEmbedding_ = tf.cast(positionEmbedding, dtype=tf.float32) # 得到三維的矩陣[batchSize, sequenceLen, embeddingSize] positionEmbedded = tf.nn.embedding_lookup(positionEmbedding_, positionIndex) return positionEmbedded
四、定義訓練、測試、預測
import sys import os BASE_DIR = os.path.dirname(os.path.abspath(__file__)) #當前程式上一級目錄，這裡為transformer from dataset.fudanDataset import FudanDataset from models.transformer import Transformer from utils.utils import * from utils.metrics import * from config.fudanConfig import FudanConfig from config.globalConfig import PATH import numpy as numpy import tensorflow as tf import time import datetime from tkinter import _flatten from sklearn import metrics import jieba def train(): print("配置Saver。。。\n") save_dir = 'checkpoint/transformer/' if not os.path.exists(save_dir): os.makedirs(save_dir) save_path = os.path.join(save_dir, 'best_validation') # 最佳驗證結果儲存路徑 globalStep = tf.Variable(0, name="globalStep", trainable=False) # 配置 Saver saver = tf.train.Saver() #定義session """ session_conf = tf.ConfigProto(allow_soft_placement=True, log_device_placement=False) session_conf.gpu_options.allow_growth=True session_conf.gpu_options.per_process_gpu_memory_fraction = 0.9 # 配置gpu佔用率 sess = tf.Session(config=session_conf) """ sess = tf.Session() print("定義優化器。。。\n") # 定義優化函式，傳入學習速率引數 optimizer = tf.train.AdamOptimizer(config.trainConfig.learningRate) # 計算梯度,得到梯度和變數 gradsAndVars = optimizer.compute_gradients(model.loss) # 將梯度應用到變數下，生成訓練器 trainOp = optimizer.apply_gradients(gradsAndVars, global_step=globalStep) sess.run(tf.global_variables_initializer()) def trainStep(batchX, batchY): """ 訓練函式 """ feed_dict = { model.inputX: batchX, model.inputY: batchY, model.dropoutKeepProb: config.modelConfig.dropoutKeepProb, } _, step, loss, predictions = sess.run([trainOp, globalStep, model.loss, model.predictions], feed_dict) if config.numClasses == 1: acc, recall, prec, f_beta = get_binary_metrics(pred_y=predictions, true_y=batchY) elif config.numClasses > 1: acc, recall, prec, f_beta = get_multi_metrics(pred_y=predictions, true_y=batchY, labels=labelList) return loss, acc, prec, recall, f_beta def valStep(batchX, batchY): """ 驗證函式 """ feed_dict = { model.inputX: batchX, model.inputY: batchY, model.dropoutKeepProb: 1.0, } step, loss, predictions = sess.run([globalStep, model.loss, model.predictions], feed_dict) if config.numClasses == 1: acc, recall, prec, f_beta = get_binary_metrics(pred_y=predictions, true_y=batchY) elif config.numClasses > 1: acc, recall, prec, f_beta = get_multi_metrics(pred_y=predictions, true_y=batchY, labels=labelList) return loss, acc, prec, recall, f_beta print("開始訓練。。。\n") best_f_beta_val = 0.0 # 最佳驗證集準確率 last_improved = 0 # 記錄上一次提升批次 require_improvement = 1000 # 如果超過1000輪未提升，提前結束訓練 flag = False for epoch in range(config.trainConfig.epoches): print('Epoch:', epoch + 1) batch_train = batch_iter(train_data, train_label, config.batchSize) for x_batch, y_batch in batch_train: loss, acc, prec, recall, f_beta = trainStep(x_batch, y_batch) currentStep = tf.train.global_step(sess, globalStep) # 多少次迭代列印一次訓練結果： if currentStep % config.trainConfig.print_per_step == 0: print("train: step: {}, loss: {:.4f}, acc: {:.4f}, recall: {:.4f}, precision: {:.4f}, f_beta: {:.4f}".format( currentStep, loss, acc, recall, prec, f_beta)) if currentStep % config.trainConfig.evaluateEvery == 0: print("開始驗證。。。\n") losses = [] accs = [] f_betas = [] precisions = [] recalls = [] batch_val = batch_iter(val_data, val_label, config.batchSize) for x_batch, y_batch in batch_val: loss, acc, precision, recall, f_beta = valStep(x_batch, y_batch) losses.append(loss) accs.append(acc) f_betas.append(f_beta) precisions.append(precision) recalls.append(recall) if mean(f_betas) > best_f_beta_val: # 儲存最好結果 best_f_beta_val = mean(f_betas) last_improved = currentStep saver.save(sess=sess, save_path=save_path) improved_str = '*' else: improved_str = '' time_str = datetime.datetime.now().isoformat() print("{}, step: {:>6}, loss: {:.4f}, acc: {:.4f},precision: {:.4f}, recall: {:.4f}, f_beta: {:.4f} {}".format( time_str, currentStep, mean(losses), mean(accs), mean(precisions), mean(recalls), mean(f_betas), improved_str)) if currentStep - last_improved > require_improvement: # 驗證集正確率長期不提升，提前結束訓練 print("沒有優化很長一段時間了，自動停止") flag = True break # 跳出迴圈 if flag: # 同上 break sess.close() def test(test_data,test_label): print("開始進行測試。。。") save_path = 'checkpoint/transformer/best_validation' saver = tf.train.Saver() sess = tf.Session() sess.run(tf.global_variables_initializer()) saver.restore(sess=sess, save_path=save_path) # 讀取儲存的模型 data_len = len(test_data) test_batchsize = 128 batch_test = batch_iter(test_data, test_label, 128, is_train=False) pred_label = [] for x_batch,y_batch in batch_test: feed_dict = { model.inputX: x_batch, model.inputY: y_batch, model.dropoutKeepProb: 1.0, } predictions = sess.run([model.predictions], feed_dict) pred_label.append(predictions[0].tolist()) pred_label = list(_flatten(pred_label)) test_label = [np.argmax(item) for item in test_label] # 評估 print("計算Precision, Recall and F1-Score...") print(metrics.classification_report(test_label, pred_label, target_names=true_labelList)) sess.close() def process_sentence(data): fudanDataset._get_stopwords() sentence_list = [] for content in data: words_list = jieba.lcut(content, cut_all=False) tmp1 = [] for word in words_list: word = word.strip() if word not in fudanDataset.stopWordDict and word != '': tmp1.append(word) else: continue sentence_list.append(tmp1) vocab = fudanDataset._get_vocaburay() word2idx,idx2word = fudanDataset._wordToIdx() label2idx,idx2label = fudanDataset._labelToIdx() res_data = [] #print(content) for content in sentence_list: tmp2 = [] if len(content) >= config.sequenceLength: #大於最大長度進行截斷 content = content[:config.sequenceLength] else: #小於最大長度用PAD的id進行填充層 content = ['PAD']*(config.sequenceLength-len(content)) + content for word in content: #將詞語用id進行對映 if word in word2idx: tmp2.append(word2idx[word]) else: tmp2.append(word2idx['UNK']) res_data.append(tmp2) return res_data def get_predict_content(content_path,label_path): use_data = 5 txt_list = [] label_list = [] predict_data = [] predict_label = [] content_file = open(content_path,"r",encoding="utf-8") label_file = open(label_path,"r",encoding="utf-8") for txt in content_file.readlines(): #讀取每一行的txt txt = txt.strip() #去除掉\n txt_list.append(txt) for label in label_file.readlines(): label = label.strip() label_list.append(label) data = [] for txt,label in zip(txt_list,label_list): data.append((txt,label)) import random predict_data = random.sample(data,use_data) p_data = [] p_label = [] for txt,label in predict_data: with open(txt,"r",encoding="gb18030",errors='ignore') as fp1: tmp = [] for line in fp1.readlines(): #讀取每一行 tmp.append(line.strip()) p_data.append("".join(tmp)) p_label.append(label) content_file.close() label_file.close() return p_data,p_label def predict(data,label,p_data): print("開始預測文字的類別。。。") predict_data = data predict_true_data = label save_path = 'checkpoint/transformer/best_validation' saver = tf.train.Saver() sess = tf.Session() sess.run(tf.global_variables_initializer()) saver.restore(sess=sess, save_path=save_path) # 讀取儲存的模型 feed_dict = { model.inputX: predict_data, model.inputY: predict_true_data, model.dropoutKeepProb: 1.0, } predictions = sess.run([model.predictions], feed_dict) pred_label = predictions[0].tolist() real_label = [np.argmax(item) for item in predict_true_data] for content,pre_label,true_label in zip(p_data,pred_label,real_label): print("輸入的文字是：{}...".format(content[:100])) print("預測的類別是：",idx2label[pre_label]) print("真實的類別是：",idx2label[true_label]) print("================================================") sess.close() if __name__ == '__main__': config = FudanConfig() fudanDataset = FudanDataset(config) word2idx,idx2word = fudanDataset._wordToIdx() label2idx,idx2label = fudanDataset._labelToIdx() print("載入資料。。。") train_content_path = os.path.join(PATH, "process/Fudan/word2vec/data/train_content.txt") train_label_path = os.path.join(PATH, "process/Fudan/train_label.txt") test_content_path = os.path.join(PATH, "process/Fudan/word2vec/data/test_content.txt") test_label_path = os.path.join(PATH, "process/Fudan/test_label.txt") fudanDataset._getTrainValData(train_content_path,train_label_path) fudanDataset._getTestData(test_content_path,test_label_path) fudanDataset._getWordEmbedding() train_data,val_data,train_label,val_label = fudanDataset.trainData,fudanDataset.valData,fudanDataset.trainLabels,fudanDataset.valLabels test_data,test_label = fudanDataset.testData,fudanDataset.testLabels train_label = one_hot(train_label) val_label = one_hot(val_label) test_label = one_hot(test_label) wordEmbedding = fudanDataset.wordEmbedding labelList = fudanDataset.labelList true_labelList = [idx2label[label] for label in labelList] print("定義模型。。。") model = Transformer(config, wordEmbedding) test(test_data,test_label) print("進行預測。。。") p_data,p_label = get_predict_content(os.path.join(PATH, "process/Fudan/test.txt"),test_label_path) process_data = process_sentence(p_data) onehot_label = np.zeros((len(p_label),config.numClasses)) for i,value in enumerate(p_label): onehot_label[i][label2idx[value]] = 1 process_label = onehot_label predict(process_data,process_label,p_data)
結果：
計算Precision, Recall and F1-Score... precision recall f1-score support Agriculture 0.83 0.90 0.87 1022 Art 0.79 0.86 0.82 742 Communication 0.00 0.00 0.00 27 Computer 0.93 0.97 0.95 1358 Economy 0.87 0.89 0.88 1601 Education 0.67 0.07 0.12 61 Electronics 0.00 0.00 0.00 28 Energy 1.00 0.03 0.06 33 Enviornment 0.86 0.95 0.90 1218 History 0.68 0.66 0.67 468 Law 0.18 0.12 0.14 52 Literature 0.00 0.00 0.00 34 Medical 0.19 0.06 0.09 53 Military 0.50 0.03 0.05 76 Mine 1.00 0.03 0.06 34 Philosophy 0.62 0.22 0.33 45 Politics 0.78 0.88 0.83 1026 Space 0.91 0.81 0.85 642 Sports 0.86 0.88 0.87 1254 Transport 1.00 0.02 0.03 59 accuracy 0.84 9833 macro avg 0.63 0.42 0.43 9833 weighted avg 0.83 0.84 0.83 9833
輸入的文字是：中國環境科學CHINA ENVIRONMENTAL SCIENCE1998年 第18卷 第1期 No.1 Vol.18 1998科技期刊鎘脅迫對小麥葉片細胞膜脂過氧化的影響*羅立新 孫鐵珩 靳月華(中... 預測的類別是： Enviornment 真實的類別是： Enviornment ================================================ 輸入的文字是：自動化學報AGTA AUTOMATICA SINICA1999年 第25卷 第2期 Vol.25 No.2 1999TSP問題分層求解演算法的複雜度研究1)盧 欣 李衍達關鍵詞 TSP，區域性搜尋演算法，動... 預測的類別是： Computer 真實的類別是： Computer ================================================ 輸入的文字是：【 文獻號 】3-5519【原文出處】人民日報【原刊地名】京【原刊期號】19960615【原刊頁號】⑵【分 類 號】D4【分 類 名】中國政治【影印期號】199606【 標 題 】中國人民政治協商會... 預測的類別是： Politics 真實的類別是： Politics ================================================ 輸入的文字是：軟體學報JOURNAL OF SOFTWARE1999年 第2期 No.2 1999視覺導航中基於模糊神經網的消陰影演算法研究郭木河 楊 磊 陶西平 何克忠 張 鈸摘要 在實際的應用中,由於室外移動機器... 預測的類別是： Computer 真實的類別是： Computer ================================================ 輸入的文字是：【 文獻號 】2-814【原文出處】中國鄉鎮企業會計【原刊地名】京【原刊期號】199907【原刊頁號】7～9【分 類 號】F22【分 類 名】鄉鎮企業與農場管理【影印期號】199908【 標 題 】... 預測的類別是： Economy 真實的類別是： Economy ================================================
總目錄結構：
後續還將繼續新增相應的功能，比如tensorboard視覺化，其它網路LSTM、GRU等等。
參考：
https://www.cnblogs.com/jiangxinyang/p/10210813.html
這裡面在transformer模型多頭注意力那裡漏了一句：
keyMasks = tf.sign(tf.abs(tf.reduce_sum(rawKeys, axis=-1)))
導致弄了好久，同時也參考了：
https://github.com/gaussic/text-classification-cnn-rnn
基本上是結合了這兩個地方的程式碼再進行了些重構。