-
Notifications
You must be signed in to change notification settings - Fork 0
/
Bert.py
379 lines (318 loc) · 17 KB
/
Bert.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
import os
import numpy as np
import pandas as pd
import tensorflow as tf
import tensorflow_hub as hub
import tensorflow_text as text
from official.nlp import optimization # to create AdamW optimizer
import matplotlib.pyplot as plt
tf.get_logger().setLevel('ERROR')
#The bert classifier is based on the implementation from:
#https://www.tensorflow.org/text/tutorials/classify_text_with_bert#about_bert
class Bert:
'''
Bert text classifiert using tensorflow and keras.
'''
def __init__(self, num_classes,bert_model_name = 'bert_en_cased_L-12_H-768_A-12', random_state = None):
'''
input:
num_classes: number of classes in the data set
bert_model_name: bert model type. default is 'bert_en_cased_L-12_H-768_A-12'
random_state: a random state for reproducebility. default is None
'''
map_name_to_handle = {
'bert_en_uncased_L-12_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_uncased_L-12_H-768_A-12/3',
'bert_en_cased_L-12_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_cased_L-12_H-768_A-12/3',
'bert_multi_cased_L-12_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_multi_cased_L-12_H-768_A-12/3',
'small_bert/bert_en_uncased_L-2_H-128_A-2':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-2_H-128_A-2/1',
'small_bert/bert_en_uncased_L-2_H-256_A-4':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-2_H-256_A-4/1',
'small_bert/bert_en_uncased_L-2_H-512_A-8':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-2_H-512_A-8/1',
'small_bert/bert_en_uncased_L-2_H-768_A-12':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-2_H-768_A-12/1',
'small_bert/bert_en_uncased_L-4_H-128_A-2':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-128_A-2/1',
'small_bert/bert_en_uncased_L-4_H-256_A-4':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-256_A-4/1',
'small_bert/bert_en_uncased_L-4_H-512_A-8':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-512_A-8/1',
'small_bert/bert_en_uncased_L-4_H-768_A-12':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-4_H-768_A-12/1',
'small_bert/bert_en_uncased_L-6_H-128_A-2':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-6_H-128_A-2/1',
'small_bert/bert_en_uncased_L-6_H-256_A-4':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-6_H-256_A-4/1',
'small_bert/bert_en_uncased_L-6_H-512_A-8':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-6_H-512_A-8/1',
'small_bert/bert_en_uncased_L-6_H-768_A-12':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-6_H-768_A-12/1',
'small_bert/bert_en_uncased_L-8_H-128_A-2':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-8_H-128_A-2/1',
'small_bert/bert_en_uncased_L-8_H-256_A-4':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-8_H-256_A-4/1',
'small_bert/bert_en_uncased_L-8_H-512_A-8':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-8_H-512_A-8/1',
'small_bert/bert_en_uncased_L-8_H-768_A-12':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-8_H-768_A-12/1',
'small_bert/bert_en_uncased_L-10_H-128_A-2':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-10_H-128_A-2/1',
'small_bert/bert_en_uncased_L-10_H-256_A-4':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-10_H-256_A-4/1',
'small_bert/bert_en_uncased_L-10_H-512_A-8':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-10_H-512_A-8/1',
'small_bert/bert_en_uncased_L-10_H-768_A-12':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-10_H-768_A-12/1',
'small_bert/bert_en_uncased_L-12_H-128_A-2':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-12_H-128_A-2/1',
'small_bert/bert_en_uncased_L-12_H-256_A-4':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-12_H-256_A-4/1',
'small_bert/bert_en_uncased_L-12_H-512_A-8':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-12_H-512_A-8/1',
'small_bert/bert_en_uncased_L-12_H-768_A-12':
'https://tfhub.dev/tensorflow/small_bert/bert_en_uncased_L-12_H-768_A-12/1',
'albert_en_base':
'https://tfhub.dev/tensorflow/albert_en_base/2',
'electra_small':
'https://tfhub.dev/google/electra_small/2',
'electra_base':
'https://tfhub.dev/google/electra_base/2',
'experts_pubmed':
'https://tfhub.dev/google/experts/bert/pubmed/2',
'experts_wiki_books':
'https://tfhub.dev/google/experts/bert/wiki_books/2',
'talking-heads_base':
'https://tfhub.dev/tensorflow/talkheads_ggelu_bert_en_base/1'}
map_model_to_preprocess = {
'bert_en_uncased_L-12_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'bert_en_cased_L-12_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_cased_preprocess/3',
'small_bert/bert_en_uncased_L-2_H-128_A-2':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-2_H-256_A-4':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-2_H-512_A-8':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-2_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-4_H-128_A-2':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-4_H-256_A-4':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-4_H-512_A-8':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-4_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-6_H-128_A-2':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-6_H-256_A-4':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-6_H-512_A-8':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-6_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-8_H-128_A-2':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-8_H-256_A-4':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-8_H-512_A-8':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-8_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-10_H-128_A-2':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-10_H-256_A-4':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-10_H-512_A-8':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-10_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-12_H-128_A-2':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-12_H-256_A-4':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-12_H-512_A-8':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'small_bert/bert_en_uncased_L-12_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'bert_multi_cased_L-12_H-768_A-12':
'https://tfhub.dev/tensorflow/bert_multi_cased_preprocess/3',
'albert_en_base':
'https://tfhub.dev/tensorflow/albert_en_preprocess/3',
'electra_small':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'electra_base':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'experts_pubmed':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'experts_wiki_books':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3',
'talking-heads_base':
'https://tfhub.dev/tensorflow/bert_en_uncased_preprocess/3'}
## set random state if given
if random_state is not None:
self.random_state = random_state
np.random.seed(random_state)
tf.random.set_seed(random_state)
## select encoder and prepocessing model.
self.tfhub_handle_encoder = map_name_to_handle[bert_model_name]
self.tfhub_handle_preprocess = map_model_to_preprocess[bert_model_name]
print(f'BERT model selected : {self.tfhub_handle_encoder}')
print(f'Preprocess model auto-selected: {self.tfhub_handle_preprocess}')
## build the model
self.model = self.__build_classifier_model(num_classes)
def __build_classifier_model(self,num_classes):
'''
helper method to build the bert classifier model
'''
text_input = tf.keras.layers.Input(shape=(), dtype=tf.string, name='text')
preprocessing_layer = hub.KerasLayer(self.tfhub_handle_preprocess, name='preprocessing')
encoder_inputs = preprocessing_layer(text_input)
encoder = hub.KerasLayer(self.tfhub_handle_encoder, trainable=True, name='BERT_encoder')
outputs = encoder(encoder_inputs)
net = outputs['pooled_output']
net = tf.keras.layers.Dropout(0.1)(net)
net = tf.keras.layers.Dense(num_classes, activation='softmax', name='classifier')(net)
return tf.keras.Model(text_input, net)
def plot_model(self):
'''
method to plot the model structure
'''
return tf.keras.utils.plot_model(self.model)
def __data_preprocessing(self,data,batch_size, training = True):
'''
helper method for preprocessing the data. Convert to correct datastructer for tensorflow
'''
## convert categorical to int. and store the int to label conversion in a dict.
if training:
self.label_map = {}
for i,label in enumerate(data.label.unique()):
self.label_map[i] = label
for number,label in self.label_map.items():
data.label[data.label==label] = number
data.label = data.label.astype("int32")
## convert to tensorflow data set
nrow = data.shape[0]
data = tf.data.Dataset.from_tensor_slices((data.text, data.label))
if self.random_state is not None:
data = data.shuffle(nrow*2,self.random_state)
else:
data = data.shuffle(nrow*2)
data = data.batch(batch_size)
data = data.cache().prefetch(buffer_size=tf.data.AUTOTUNE)
return data
def from_path_data_preprocessing(self,path,batch_size,training = True):
'''
makes a data from a csv file ready for training by converting it to a tensorflow dataset
'''
data = pd.read_csv(path)
return self.__data_preprocessing(data,batch_size,training=training)
def data_preprocessing(self,X,y,batch_size,training = True):
'''
merges the text and label into a data frame makes ready for training by converting it to a tensorflow dataset
input:
X: the text array
y: the label array
batch_size: the size of the batches when training
training: whether the method is used when traning the model or not
'''
if isinstance(X, np.ndarray):
X =pd.DataFrame(X)
if isinstance(y, np.ndarray):
y =pd.DataFrame(y)
data = pd.concat([X,y],axis=1)
data.columns = ["text","label"]
return self.__data_preprocessing(data,batch_size,training = training)
def __train(self,data,learning_rate,batch_size = 64,epochs=10):
'''
helper method. Training the model.
'''
## loss
loss = tf.keras.losses.SparseCategoricalCrossentropy()
metrics = tf.metrics.SparseCategoricalAccuracy('accuracy')
#optimizer
steps_per_epoch = tf.data.experimental.cardinality(data).numpy()
num_train_steps = steps_per_epoch * epochs
num_warmup_steps = int(0.1*num_train_steps)
optimizer = optimization.create_optimizer(init_lr=learning_rate,
num_train_steps=num_train_steps,
num_warmup_steps=num_warmup_steps,
optimizer_type='adamw')
self.model.compile(optimizer=optimizer,loss=loss,metrics=metrics)
return self.model.fit(x=data,epochs=epochs)
def train(self,X,y, learning_rate,batch_size = 64,epochs=10):
'''
Training the model.
input:
X: the text array
y: the label array
learning_rate: the learning rate used when training
batch_size: the size of the batches when training
epochs: the number of epochs to train
'''
train_ds = self.data_preprocessing(X,y,batch_size = batch_size)
self.__train(train_ds,learning_rate,batch_size = batch_size,epochs=epochs)
def train_from_path(self,path, learning_rate,batch_size = 64, epochs=10):
'''
Training the model using the file path of the training data
input:
path: the file path to the traning data.
learning_rate: the learning rate used when training
batch_size: the size of the batches when training
epochs: the number of epochs to train
'''
train_ds = self.from_path_data_preprocessing (path=path,batch_size = batch_size)
self.__train(train_ds,learning_rate,batch_size = batch_size,epochs=epochs)
def __predict_proba(self,X, batch_size=1):
'''
helper method for finding the probability of the predictions
'''
X = tf.data.Dataset.from_tensor_slices(X)
X = X.batch(batch_size)
X = X.cache().prefetch(buffer_size=tf.data.AUTOTUNE)
return self.model.predict(X)
def predict_label_proba(self,X, batch_size=1):
'''
find the the label prediction and the corresponding prediction probability
input:
X: pandas dataframe containing the text or a list of text
'''
y_pred = self.__predict_proba(X,batch_size=batch_size)
prop = y_pred.max(axis=-1)
lab = np.array(list(map(lambda x:self.label_map[x] ,y_pred.argmax(axis=-1))))
return list(zip(lab,prop))
def predict(self,X, batch_size=1):
'''
returns the the label prediction of the input text
input:
X: pandas dataframe containing the text or a list of text
'''
return self.__predict_proba(X,batch_size=batch_size).argmax(axis=-1)
def evaluate_from_path(self,path):
'''
evaluetes the model on the test data using the file path to the test data
input:
path: the file path to the test data
'''
test = self.from_path_data_preprocessing (path=path,batch_size = 1, training = False)
return self.model.evaluate(test)
def evaluate(self,data):
'''
evaluetes the model on the test data.
input:
data: the test data, pandas data frame
'''
test = self.__data_preprocessing(data,batch_size=1, training = False)
return self.model.evaluate(test)
if __name__ == "__main__":
batch_size = 2
seed = 42
bert = Bert(num_classes = 2, random_state = seed)
bert.train_from_path("/data/train_labeled.csv",learning_rate= 0.001,batch_size = batch_size,epochs = 2)
test = pd.read_csv("/data/test.csv")
print(bert.predict_label_proba(test.text[:10]))