三步使用bert搭建文本分类器

最新推荐文章于 2024-07-22 10:22:43 发布

xienan8

最新推荐文章于 2024-07-22 10:22:43 发布

阅读量948

点赞数

分类专栏： NLP

本文链接：https://blog.csdn.net/xienan8/article/details/117369168

版权

NLP 专栏收录该内容

2 篇文章 1 订阅

订阅专栏

不说废话，直接三步搭建最简单的bert文本多标签分类器

1.去官网https://github.com/google-research/bert 下载一个bert模型

2.搭建bert-service https://github.com/hanxiao/bert-as-service

3.分类demo

mb = MultiLabelBinarizer()
dataset = pd.read_csv('train.csv')
x = dataset['corpus']
y = [_.split("&&&&") for _ in dataset['label']]
y = mb.fit_transform(np.array(y))
x_train, x_test, y_train, y_test = train_test_split(x, y)
bc = BertClient()
X_train = bc.encode(x_train)
X_test = bc.encode(x_test)
model = Sequential()
model.add(Dense(100, activation='relu', input_dim=768))
model.add(Dropout(0.5))
model.add(Dense(5, activation='sigmoid'))
model.compile(loss="binary_crossentropy", optimizer='adam', metrics=['accuracy'])
H = model.fit(X_train, y_train, epochs=15, validation_split=0.2)
print(H.history())
predictions = model.predict(X_test)
predictions = predictions.argmax(axis=1)
print(classification_report(y_test.argmax(axis=1), predictions))

end

！！！！！！！！！！！！！！！！！！！！！！